Research proposals

PhD selection is performed twice a year through an open process. Further information are available at http://dottorato.polito.it/en/admission.

Grants from Politecnico di Torino and other bodies are available.

In the following, we report the list of topics for new PhD students in Computer and Control Engineering. This list will be regularly updated when new proposals will be available.

If you are interested in one of the topics, please contact the related Proposer.


- Design of Secure Computer Hardware Architectures
- AI-oriented Design Methods for Intelligent Aerospace Computing Systems
- Computational Intelligence for Computer Aided Design
- Anomaly detection and knowledge extraction in household energy consumption
- Automatic configuration techniques for improving the exploitation of the emergin...
- Design of a framework for supporting the development of new computational paradi...
- Exploring the use of Deep Natural Language Processing models to analyze document...
- Analysis, search, development and reuse of smart contracts
- Dependable Computing in the heterogenous era: Fault Tolerance for the Middle Lay...
- Beyond 5G (B5G) and the road to 6G Networks
- Media Quality Optimization using Machine Learning on Large Scale Datasets
- Digital Wellbeing in the Internet of Things
- Advancing Mobile UI Testing Techniques
- Cybersecurity applied to embedded control systems
- Planning Safety in the Era of Approximate Computing Systems
- Exploring Deep Learning for In-Field Fault Detection in Cyber Physical Systems
- Testing and Reliability of Automotive oriented System-on-Chip
- Graph network models for Data Science
- Promoting Diversity in Evolutionary Algorithms
- Accessibility of Internet of Things Devices and Systems
- Automatic Generation of biofabrication pProtocols for regenerative medicine appl...
- Pervasive Information Management
- Integration of Machine Learning and Network Virtualization for the Orchestration...
- Architectures and Protocols for the Management of the IoT – Edge Computing Eco...
- Local energy markets in citizen-centred energy communities
- Intermittent Computing for Batteryless Systems
- Automatic hardware-aware design and optimization of deep learning models
- Simulation and Digital Twin development for Industrial Cyber Physical Systems
- Algorithms, architectures and technologies for ubiquitous applications
- Gamification of E2E Software Testing
- Stimulating the Senses in Virtual Environments
- Context and Emotion Aware Embodied Conversational Agents
- Robotic 3D Vision for Scene Understanding Across Domains
- Proposal title (max 150 characters) Automated Painting: Object Modeli...
- Convex relaxation based methods for gray-box model identification
- Techniques for rapid deployment of dependable automotive computing architectures...
- When the cloud meets the edge: Reusing your available resources with opportunist...
- Liquid computing: orchestrating the computing continuum
- Attention-guided cross domain visual geo-localization
- Unsupervised cross domain detection and retrieval from scarce data for monitorin...
- Cross Modal Neural Architecture Search
- Generative modelling for improved decision making in data-limited applications ...
- PoliTO-EURECOM PhD on methods for modelling software protection as a risk analys...
- Combining Meta-Learning and Self-Supervision for Test-Time Adaptation
- Autonomous systems' reliability and security
- Human-Centered AI for Smart Environments
- Vehicular micro clouds in 5G edge/cloud infrastructures
- Urban intelligence
- Explainable AI (XAI)
- Quantum Machine Learning Applications
- Virtual and Augmented Reality: new frontiers in education and training
- Web-based distributed real-time sonic interaction
- Sparse optimization for system identification and machine learning
- ICT technologies for the automated assessment and management of neurological dis...
- Simulation and Modelling of Energy Demand Flexibility in energy communities
- Co-simulation platform for real-time analysis of smart energy communities
- Cybersecurity Automation for Cyber-Physical Systems
- Reliability and Safety of Neural Networks running on GPUs
- Zero-trust security of network nodes
- Cybersecurity in the Quantum Era
- Evaluating the impact of automated decision systems in urban contexts


Detailed descriptions

Title: Design of Secure Computer Hardware Architectures
Proposer: Paolo Prinetto
Group website: https://www.testgroup.polito.it
Summary of the proposal: In recent times, security has taken on the role of essential requirement for any type of computing machine, from the server clusters of large corporates to end-points represented by smart objects surrounding our daily lives. The communication encryption systems, which are considered the basis against the misuse of data and functions, are fundamental but not sufficient. Computing systems need to implement defenses that are not limited to application data protection, but scale more and more towards the lowest levels of abstraction, until they reach the hardware on which programs and services run. If the concepts of virtualization, isolation, supervision, memory protection and secure execution are included in the design paradigm of microprocessors and hardware components, the systems would benefit from an architectural protection that goes beyond all possible vulnerabilities of software or communication protocols.
The PhD proposals aims at proposing, studying, and developing architectural solutions for computing hardware, able to guarantee predefined level of security to systems. The work will target examples of open core architecture suitable for research purposes (e.g., RISC-V) to reason about, elaborate and test such solutions.
The topic is so huge that two candidates are foreseen.
Rsearch objectives and methods: The research objectives will be:
1) Analyzing security issues by considering different vulnerabilities and possible attack surfaces. Vulnerabilities can be generated by the hardware itself or by the software it runs. A hardware vulnerability can be exploited by a combination of machine instructions that are executed on the target, and has several malicious implications: private information extracted from the cache by means of memory response time measurements, improper access to privileged processor modes, artificial creation of perpetual stalls of the machine that inhibit its functions, etc. A software vulnerability is instead introduced by the programmer, but if the hardware is not designed to withstand the most frequent errors, it faces a series of risks, including reading of private data of the kernel or of other users, corruption of data in memory, code injection or execution hijacking, guest system bypass, improper access to the host/supervisor system, and so on.

2) Several different solutions will be considered, depending on the type of vulnerability and the level of criticality of the system that uses the devices. These will include solutions:
a. preventing the leakage of information between different security domains, defined and enforced in hardware;
b. preventing the hijacking of the behavior of the system into malicious actions (e.g., Arbitrary Code Execution);
c. to detect potentially suspect behavior and warn the execution environment about possible threats;
d. aimed at confining the potential malicious execution outside the access to critical resources, without affecting the overall execution environment;
e. exploiting trust zones, i.e., regions of the computing hardware used to run critical code or to store critical information;
f. exploiting coordination between existing features of processors to implement security policies.

While the results obtained during the research period are expected to be general and hold for any platform, the work during the PhD thesis will explicitly focus on the RISC-V open architecture (https://riscv.org/).
Outline of work plan: The following steps will be taken:

1. Comprehensive analysis of:

o vulnerabilities and attacks
o proposed hardware-based countermeasures
o their effectiveness;

2. Identifying a proper set of case studies;

3. Designing a set of solutions exploiting the alternatives summarized at point 2) of the previous section on Research objectives;

4. Developing, comparing, evaluating, gathering experimental results from the implementation for the different solutions;

5. Leveraging on the acquired experience, defining:
o a set of rules to be adopted to design secure hardware architectures;
o Design-for-Trust and Design-for-Security rules to be followed when designing new systems.

The candidates are expected to perform:
• steps 1-2 in the 1st year,
• steps 3-4 in the 2nd year,
• steps 5 in the 3rd year.
Expected target publications: Conferences:
- IEEE European Symposium on Security and Privacy (EuroSandP)
- Design, Automation and Test in Europe (DATE)
- VLSI Test Symposium (VTS)
- European Test Symposium (ETS)
- HOST
- Specialized Workshops
- ITASEC

Journals:
- IEEE Transactions on Dependable and Secure Computing
- IEEE Transactions on Computers
- IEEE Internet of Things
- ACM Transactions on Information and System Security
- ACM Transactions on Privacy and Security
- Design and Test of Computers
Current funded projects of the proposer related to the proposal:
Possibly involved industries/companies:1. CINI Cybersecurity National Lab
2. Fondazione Links
3. Blu5 Labs

Title: AI-oriented Design Methods for Intelligent Aerospace Computing Systems
Proposer: Luca Sterpone
Group website: http://www.cad.polito.it
Summary of the proposal: The revolution of modern computing systems has been widely supported by advancements in computer systems architecture and hardware technologies. Today, computing systems and electronic circuits may play a fundamental role in the automatization of aerospace missions in 3 to 5 years in the future. On one side, the automatization process is central in reducing the costs of space missions by increasing the performances and the efficiency of satellites; on the other side, the adoption of Artificial Intelligence (AI) on satellite computing systems will enable to take autonomous decisions independently from the ground control with enormous advantages for the space mission goals. Some initial tentative of adopting AI in the satellite has been already successfully applied to earth observation satellites, however, the overall AI-oriented computing approach is far from being widely used in space due to the limited knowledge on the behavior of AI algorithm in the harsh space environment.
The aim of the present research proposal targets the pipeline stages of an AI algorithm in space considering the design, development, and final deployment phases. The activities performed in the research proposal will cover two directions. The first will target Computer Aided Design (CAD) compilation and synthesis methods for designing resilient and efficient AI algorithms, while the latter will target the robust mapping of AI algorithms on the modern generation of edge devices such as NVIDIA Jetson, Xilinx AP-SoC, and Intel Stratix.
Rsearch objectives and methods: A successful realization of AI algorithms in aerospace computing systems requires a drastic improvement of the available CAD tools. The strategic objective of this research proposal is to improve the nowadays CAD tool approaches towards the realization of resilient and efficient AI-algorithms on computing systems using edge technologies. The main objectives are:

1. Identify the critical points of AI algorithms considering the reliability and efficiency aspects in the context of the harsh environmental conditions specific for space missions. The objective will consider Low Earth Orbit (LEO), Geostationary orbit (GEO), and Deep-Space mission conditions. These general space mission conditions will provide realistic constraints for the energy available for the computing system and the reliability specifically affected by the failure impacts on the hardware computing system that will undergo during the mission.

2. Create an AI-oriented compilation and synthesis model able to cope with the energy and reliability constraints and to eventually reduce the energy required and minimize the impact of induced failures on the overall reliability of the computing system. This objective will be supported by High-Level Synthesis (HLS), TensorFlow, and PyTorch gears. The creation of the model will be applied to a set of software algorithms adopted in the framework of the Milani mission Flight Computer (HERA Mission) and with respect to state-of-the-art methods adopted in the field of computer visions such as stereo depth estimation and binary classifications.

3. Physical deployment of AI algorithms on computing systems for a space mission. This objective will target physical design problems such as routability and timing-driven placement of resources applied to reconfigurable devices such as Xilinx AP-SoC or GPU accelerators such as NVIDIA Jetson. The physical implementation will enforce the design learning processes for the routing strategies since the proposed method will integrate Machine Learning (ML) models for the routability and timing prediction. Besides, the physical implementation will enforce specific fault tolerance solutions suitable for transients and permanent effects induced by the aerospace environment on the electronic system parts.
Outline of work plan: The workplan of the proposal is organized in the following phases:

1. Modeling of the Space Environment conditions within AI state-of-the-art tools. The state-of-the-art AI tools for the design of autonomous space applications will be explored and modeled with respect to the space environment conditions. The space environment will deeply affect the computing correctness and efficiency.

2. Performance and Reliability Degradation of the AI computation. The computational characteristics of AI algorithms will be evaluated by developing ad-hoc injection methods at different abstraction levels by adopting a top-down approach: from Python models to hardware implemented algorithms. The expected outcome is the identification of critical behavior and the definition of analysis and hardening solutions. A specific effort will be done in the field of dynamic routing mitigation solutions for artificial neural networks.

3. Compilation and Synthesis of AI algorithms. The activity will investigate compilation and synthesis strategies applied to AI algorithms. Existing HLS approaches focusing in particular on the generation of deep neural networks will be considered. Developed approaches will be compared considering the resulting network quality, efficiency, and robustness advantages. The outcome of this phase will consist of a set of hammers and bridges for the conventional CAD flow.

4. Design of Flight Computer and Vision Algorithms. This phase will apply the developed CAD methods for Flight Computer algorithms related to the Milani cube-sat and on vision algorithms embedding Convolutional Neural Networks (CNN) and stereo-vision algorithms with binary classification. The algorithms will be implemented on commercial processors such as ARM and desktop architectures, mobile GPUs, and FPGA accelerators.

5. Experimental validation and verification. The overall research activities will be supported by experimental validation and verification adopting traditional simulation tools and developing ad-hoc simulation and emulation fault injection methods. Radiation test campaigns, supported by the active collaboration with CERN in the framework of the RADNEXT EU project proposal, will be performed to emulate the space environment are planned to prove and validate the CAD methods developed during the proposal.

Please note that the research proposal is based on the usage of NVIDIA Jetson platforms, Xilinx AP-SoC and Intel Stratix FPGAs already available within the research group.
Expected target publications: IEEE Transactions on Computers
IEEE Transactions on Software Engineering
IEEE Systems
IEEE Transactions on Aerospace and Electronic Systems
IEEE Transactions on Emerging Topics in Computing
IEEE Transactions on VLSI
Microelectronic Reliability
IEEE Transactions on Reliability
IEEE Transactions on CAD
ACM Transactions on Reconfigurable Technology and Systems
IEEE DATE
IEEE ITC
Current funded projects of the proposer related to the proposal: Tyvak International in the framework of the HERA mission for the European Space Agency
Possibly involved industries/companies:Tyvak International
ESA
OHB
INFN

Title: Computational Intelligence for Computer Aided Design
Proposer: Giovanni Squillero, Riccardo Cantoro
Group website: http://www.cad.polito.it/
Summary of the proposal: Computational Intelligence (CI) is playing an increasingly important role in many industrial sectors. Techniques ascribable to this large framework have long been used in the Computer Aided Design field: probabilistic methods for the analysis of failures or the classification of processes; evolutionary algorithms for the generation of tests; bio-inspired techniques, such as simulated annealing, for the optimization of parameters or the definition of surrogate models. The recent fortune of the term “Machine Learning” renewed the interests in many automatic processes; moreover, the publicized successes of (deep) neural networks smoothed down the bias against other non-explicable black-box approaches, such as Evolutionary Algorithms, or the use of complex kernels in linear models. The proposal would focus on the use and development of “intelligent” algorithms specifically tweaked on the need and peculiarities of CAD industries.
Rsearch objectives and methods: The goal of the research is twofold: from an academic point of view, tweaking existing methodologies, as well as developing new ones, specifically able to tackle CAD problems; from an industrial point of view, creating a highly qualified expert able to bring the scientific know-how into a company, while being also able to understand the practical needs, such as how data are selected and possibly collected. The need to team the experts from industry with more mathematically minded researchers is apparent: frequently a great knowledge of the practicalities is not accompanied by the adequate understanding of the statistical models used for analysis and predictions.

In more details, the research will almost certainly include bio-inspired techniques for generating, optimizing, minimizing test programs; statistical methods for analyzing and predicting the outcome of industrial processes (e.g., predicting the maximum operating frequency of a programmable unit based on the frequencies measured by some ring oscillators; detecting dangerous elements in a circuit; predicting catastrophic events). The activity is also like to exploit (deep) neural networks, however developing novel, creative results in this area is not a priority. On the contrary, the research shall face problem related to dimensionality reduction, feature extraction and prototypes identification/creation.
Outline of work plan: The research would start by analyzing a current practical need, namely: “predictive maintenance”. A significant amount of data is currently collected by many industries, although in a rather disorganized way. The student would start by analyzing the practical problems of data collection, storage, and transmission, while, at the same time, practicing with the principles of data profiling, classification, and regression (all topics that are currently considered part of “machine learning”). The analysis of sequences to predict the final event, or rather identify a trigger, is an open research topic, with implication far beyond CAD. Unfortunately, unlikely popular ML scenarios, the availability of data is a significant limitation, a situation sometimes labeled “small data”.

Then the research shall focus on the study of surrogate measures, that is, the use of measures that can be easily and inexpensively gathered as a proxy for others, more industrially relevant but expensive. In this regard, Riccardo Cantoro and Squillero are working with a semiconductor manufacturer for using in-situ sensors values as a proxy for the prediction of operating frequency, and they jointly supervised master students.

In general, the research would need to consider techniques less able to process large amount of information, but perhaps more “intelligent”, and to use all problem-specific knowledge available. The thesis could then proceed by tackling problems related to “dimensionality reduction”, useful to limit the number of input data of the model, and “feature selection”, essential when each single feature is the result of a costly measurement. At the same time, the research is likely to help the introduction of more advanced optimization techniques in everyday tasks.
Expected target publications: Top journals with impact factors
- ASOC – Applied Soft Computing
- TEC – IEEE Transactions on Evolutionary Computation
- TC – IEEE Transactions on Computers
Top conferences
- ITC – International Test Conference
- DATE – Design, Automation and Test in Europe Conference
- GECCO – Genetic and Evolutionary Computation Conference
- CEC/WCCI – World Congress on Computational Intelligence
- PPSN - Parallel Problem Solving From Nature
Current funded projects of the proposer related to the proposal: - The proposer is collaborating with Infineon on the subjects listed in the proposal: Two contracts have been signed, the third extension is currently under discussion; A joint paper has been published at ITC, other one was submitted, others are in preparation.
- The proposer collaborated with SPEA under the umbrella contract “Colibri”. Such contract is likely to be renewed on precisely the topics listed in the proposal.
Possibly involved industries/companies:The CAD Group has a long record of successful applications of intelligent systems in several different domains. For the specific activities, the list of possibly involved companies include: SPEA, Infineon, ST Microelectronics, Comau (through the Ph.D. student Eliana Giovannitti)

Title: Anomaly detection and knowledge extraction in household energy consumption
Proposer: Enrico Macii, Edoardo Patti
Group website: http://eda.polito.it/
Summary of the proposal: The recent technological advances in pattern recognition, machine learning and data analytics disrupted several fields of modern society. Among them, the energy sector is one of the most promising, in which new technologies would surely benefit from the insights deriving from the application of such techniques to household energy consumptions.
This research proposal concerns on developing (semi-) unsupervised models, focusing on household appliances load profiles. While in use, each appliance leaves a digital footprint behind, which embeds valuable information, such as programs, energy class labels and potential failures of the system. Accessing this precious knowledge would lead to enable a number of new services and business models that would revolutionize the future energy marketplace, e.g. energy consumption forecast, Demand Response, Demand Side Management and State Estimation.
On the one hand, such models can simultaneously target two goals: anomaly detection and knowledge extraction. This information is crucial to define the energy consumption patterns of customers improving the quality of further algorithms for smart grid management. Moreover, such kind of information could enhance the awareness of the final customers about their consumption, entailing fewer waste of energy.
On the other hand, such models can generate realistic synthetic load profiles for different appliances based on the learnt knowledge (e.g. appliance's usage, energy class etc.).
Rsearch objectives and methods: The objectives of the Ph.D. plan are the following:
1. Developing the competences to design (semi-) unsupervised models to derive useful information from household load consumption
2. Providing a scalable solution to be applied to the common household appliances either energivorous or not.
3. Presenting a scalable and easy-to-apply methodology to generate realistic and synthetic load profiles of heterogeneous appliances in terms of type, use and energy class.
4. Providing final results with a limited (i.e., acceptable) error with respect to baseline data.

The final outcome of these algorithms can be exploited to
- accurately forecast load consumption profiles of users,
- evaluate new control policies for smart grid management (e.g. Demand Response, Demand Side Managements and State Estimation),
- to feed simulation engines to realistically mimic the behavior of energy distribution networks,
- to realistically evaluate new business models and applications even considering the spread of new and more efficient appliances,
- to provide personalized tips and suggestions to customers even based on the anomaly behaviors of their appliances.

The aforementioned research activities will focus on the following main area of application in computer science: pattern recognition, machine learning and data analytics.
Outline of work plan: 1st year. The candidate will study state-of the-art techniques to i) derive information and meta-data in household energy consumptions, e.g. programs, energy class labels of appliances and ii) detect potential failures and anomalies.
At the beginning, various techniques should be considered, to classify and cluster different load energy consumption patterns of household appliances in order to prove the effectiveness of the basic approach. In this phase, the most energivorous appliances will be considered (e.g. washing machines, dishwashers and fridges).

2nd year. Based on the outcomes of the first year, the developed algorithms will be extended i) to cover more type of appliances, even more energy efficient, and ii) to discern across different possible anomalies for each appliance taken under consideration.

3rd year. The methodology and the algorithms developed in the previous years will be validated to prove their robustness and scalability in being applied on a large set of different possible appliance load profiles.
Expected target publications: IEEE Transactions on Smart Grids
IEEE Transactions on Industrial Informatics
IEEE Transactions on Emerging Topics on Computing
IEEE Transactions on Computers
IEEE Systems Journal
Pattern Recognition
Expert systems with applications
Current funded projects of the proposer related to the proposal: This research proposal concerns the development of (semi-) unsupervised models to analyze load profiles of household’s appliances. This is also one of the main objectives of the Arrowhead-tools project (European investment for Digitalization and Automation Leadership)
Possibly involved industries/companies:Midori

Title: Automatic configuration techniques for improving the exploitation of the emerging technologies in the embedded systems field
Proposer: Enrico Macii, Gianvito Urgese
Group website: https://eda.polito.it/
Summary of the proposal: Today, the process of task mapping on the hardware units available on the heterogeneous embedded systems (multi-core, CPU, GPU, FPGA) is one of the main challenges for a software developer. Several research teams are involved in the design of new solutions for enhancing compilers and programming models such as to improve the exploitation of heterogeneous architectures in the domain of the edge computing, also by supporting automatic resource allocation and optimisation procedures.
Alongside, the adoption of heterogeneous embedded systems for the analysis of sequential data streams is increasing in our industrial applications. However, the cost and complexity of software development and the energy footprint of the created solutions are still not well balanced.

The basic idea of this proposal is the definition end the design of an integrated methodology capable of decomposing data-stream applications, defined by the user in high-level programming languages, into basic atomic tasks (kernels) that can be placed on the properly-selected devices of the heterogeneous embedded system.
This automatic partitioning and allocation system will support the compiling procedure of the code for the heterogeneous architectures allowing also not overly specialized embedded programmers to fully exploit the advantages of this class of architectures.
Rsearch objectives and methods: The objectives of the PhD plan are the following:

1. Develop the competence to analyse available data from product documentation and experiments, for extraction of features in complex components and systems.

2. Analyse the state-of-the-art of the compiler technology available for heterogeneous HW architectures developed for the embedded system field of applications.

3. Develop a general (machine learning based) approach for partitioning a sequential data stream application, defined by the user in high-level programming languages, in elementary computation tiles called kernels that can be placed on the properly-selected devices of a target heterogeneous architecture.

4. Design a reliable methodology for placing the elementary kernels on the devices available on the target heterogeneous embedded systems, together with the generation of the inter-task communication interfaces.

5. Design of proof-of-concept experiments for demonstrating that the developed partitioning and allocation methodology succeed in better exploiting resources of heterogeneous HW by reducing the execution time and/or the power consumption of an application without the explicit instrumentation of the code by specialized embedded programmers.

6. Provide a framework for configuring heterogeneous embedded systems in an automatic way, so as to further facilitate the optimised porting of applications on the many emerging embedded system architectures.

The activities of research mentioned above will focus on three main areas of application:
- Video surveillance and object recognition;
- Smart energy system;
- Medical and biological data stream analysis.
Outline of work plan: 1st year. The candidate will study state-of-the-art methodologies, APIs, and frameworks used for configuring embedded system platforms. Moreover, she/he will improve skills in the machine learning techniques domain for generating automatic profiling systems capable of scanning applications written in high-level programming language and identifying the kernels (sub-optimal partitioning of instructions) to be mapped on the different devices of a heterogeneous platform. In the beginning period, various techniques should be considered in order to achieve the first results and for proving the effectiveness of the basic approaches.

2nd year. The candidate will develop a methodological approach for modelling systems and processes accordingly to the experiences obtained during the first year of research.
The basic structure of a user-friendly task partitioning and allocation tool should be developed and tested at least in one area of application such as object recognition or data-stream analysis.
The tool will have a modular structure, integrated with available technologies for defining and exploring program representations for machine learning on source code tasks at the intermediate representation level (LLVM and GIMPLE).
The task partitioning and allocation tool will be profiled and preliminary results will be produced for discussing the capacity of the tool to improve the power and/or computational performances of a set of test-bench applications configured automatically for a pool of selected target heterogeneous embedded platforms.

3rd year. The candidate will apply the proposed approach to different embedded system architectures making possible a greater generalisation of the methodology to many target heterogeneous platforms adopted in the edge-computing field for implementing AI and data-stream applications. A stable version of the automatic configuration framework will be made available for the programmers of the embedded system community.

Research activities will be carried out, partly, in collaboration with the academic partners of the Arrowhead-Tools consortium and will involve industry (STMicroelectronics).
Expected target publications: The main outcome of the project will be disseminated in three international conference papers and in at least one publication in a journal of the field.
In the following the possible conference and journal targets:
- IEEE/ACM International Conferences (e.g., DAC, DATE, AICAS, NICE, ISLPED, GLSVLSI, PATMOS, ISCAS, VLSI-SoC);
- IEEE/ACM Journals (e.g., TCAD, TETC, TVLSI, TCAS-I, TCAS-II, TCOMP), MDPI Journals (e.g., Electronics).
Current funded projects of the proposer related to the proposal: ECSEL Arrowhead-Tools
ECSEL MadeIn4
Possibly involved industries/companies:STMicroelectronics

Title: Design of a framework for supporting the development of new computational paradigms capable of exploiting neuromorphic hardware architectures
Proposer: Enrico Macii, Gianvito Urgese
Group website: https://eda.polito.it/
Summary of the proposal: Although initially intended for brain simulations, the adoption of the emerging neuromorphic hardware architectures is also appealing in fields such as IoT edge devices, high-performance computing, and robotics.
It has been proved that neuromorphic platforms provide better scalability than traditional multi-core architectures and are well suitable for the class of problems that require massive parallelism as well as the exchange of small messages for which the neuromorphic hardware has a native optimised support. Moreover, since brain-inspired, the neuromorphic technologies are identified from the scientific community as particularly adapt for low power and adaptive applications required to analyse data in real-time.
However, the tools currently available in this field are still weak and miss many useful features required to support the spreading of a new neuromorphic-based computational paradigm.

The basic idea of this proposal is the definition, and the design of a high-level framework that collects simple neuromorphic models (SNM) designed to performs small general-purpose tasks compatible with the neuromorphic hardware.
The SNM framework, once included in a user-friendly EDA tool, can be directly used by most users for describing their complex application that can be then easily executed on a neuromorphic platform.
Rsearch objectives and methods: The objectives of the PhD plan are the following:

1. Develop the competence to analyse available data from product documentation and experiments, for extraction of features in complex components and systems.

2. Evaluate the potentiality of a spiking neural network (SNN), efficiently simulated on the neuromorphic platforms, when customised at the abstraction level of a flow-graph and used for implementing a general-purpose algorithm.

3. Present a general approach for generating simplified neuromorphic models implementing basic kernels that can be exploited directly by users for describing their algorithms. The abstraction level of the models will depend on the availability of software libraries supporting the neuromorphic target hardware.

4. Design a couple of proof-of-concept applications generated by combining a set of neuromorphic models which will provide outputs with a limited (i.e., acceptable) error with respect to experimental data generated by applications implemented for general-purpose architectures. Furthermore, they should also reduce the execution time and/or the power consumption.

5. Providing a framework for generating and connecting neuromorphic models in a semi-automatic way, so as to further facilitate the modelling process and the exploration of new neuromorphic-based computational paradigms.

The activities of research above will focus on the implementation of algorithms in three main areas of application:
- Video surveillance and object recognition;
- Real-time data analysis from Industrial applications.
- Medical and biological data stream analysis and pattern matching.
Outline of work plan: 1st year. The candidate will study state-of-the-art modelling techniques that can be adapted for the neuromorphic framework and will improve the skills for generating models at different levels of abstraction.
In the beginning, various ways of modelling should be considered, such as behavioural, physical, and SNN-based in order to achieve first results and for proving the effectiveness of the basic approach for the algorithm descriptions.

2nd year. The candidate will develop a methodological and integrated approach for modelling applications and systems, accordingly to the experiences obtained during the first year of research in a multi-scenario analysis.
The basic structure of a user-friendly framework supporting the new neuromorphic computational paradigm should be developed for at least one area of application such as object recognition and pattern matching.

3rd year. The candidate will apply the proposed approach to different complex systems making possible a greater generalisation of the methodology to different domains: for instance, from the analysis of future investments in the field of neuromorphic compilers that will enhance the usability of the new generations of neuromorphic hardware soon available on the market alongside the general-purpose computing units.

The research activities will be carried out, in collaboration with the Human Brain Project partners @UMAN.
Expected target publications: The main outcome of the project will be disseminated in three international conference papers and in at least one publication in a journal of the neuromorphic field.
In the following the possible conference and journal targets:
- IEEE/ACM International Conferences (e.g., DAC, DATE, AICAS, NICE, ISLPED, GLSVLSI, PATMOS, ISCAS, VLSI-SoC);
- IEEE/ACM Journals (e.g., TCAD, TETC, TVLSI, TCAS-I, TCAS-II, TCOMP), MDPI Journals (e.g., Electronics).
Current funded projects of the proposer related to the proposal: H2020 Human Brain Project
Possibly involved industries/companies:

Title: Exploring the use of Deep Natural Language Processing models to analyze documents in cross-lingual and multi-domain scenarios
Proposer: Luca Cagliero
Group website: https://dbdmg.polito.it
Summary of the proposal: Deep Natural Language Processing (Deep NLP) focuses on applying both representation learning and Deep Learning to tackle NLP tasks. The main goal is to analyze large document corpora to support knowledge discovery. Examples of Deep NLP applications encompass extractive text summarization (i.e., generate concise summaries of large corpora), machine translation (i.e., translate a text from one language to another), sentiment analysis (i.e., extract subjective information from a text such as opinions, moods, rates, and feelings), and question answering (e.g., chatbot applications).

Vector representations of text are commonly used to capture semantic text relationships at the word-level (e.g., Word2Vect, GloVe, FastText) or at the sentence-level (e.g., BERT, XLNET, ElMO). However, existing models show a limited portability towards different languages and domains.

The Ph.D. candidate will study, design, and develop new DNLP models tailored to cross-lingual and cross-domain scenarios. Furthermore, she/he will apply the aforesaid models to solve specific NLP tasks such as cross-lingual and multi-domain text summarization and sentiment analysis.
Rsearch objectives and methods: The main goal of Deep Natural Language Processing (DNLP) techniques is to exploit Deep Learning models in order to extract significant information from large document corpora. The idea behind is to transform the original textual units into high-dimensional vector representations, which incorporate most semantic text relationships.

The current challenges are enumerated below.

- The need for large training datasets. Deep Learning architectures usually require sufficiently large amounts of training data. When there is a lack of training data, it is necessary to opportunistically reuse pre-trained models and data from other related domains. The open question is to what extent these models and datasets are suitable and for which application contexts.

- The need to cope with multilingual documents. Documents can be written in different languages. Models trained for specific languages are not always portable to other languages. This calls for new cross-lingual NLP solutions.

- The need to cope with documents relative to different domains. Documents often range over specific domains. The covered topic/domain influences text meaning, syntax, and scope. Effectively coping with multi-domain document corpora is another open research challenge. It is particularly relevant to extract meaningful document summaries and to infer the main author's opinions/feelings).

- The need to generate new text. In specific NLP applications extracting part of the existing textual content is not sufficient. For example, chatbot applications usually need to generate appropriate answers to end-user queries. The study of DNLP techniques to produce abstractive summaries of large documents is another challenging yet appealing research direction.

The main goal of the proposal is to study, develop, and test new DNLP solutions tailored to multilingual and multidomain contexts, paying a particular attention to contexts in which there is a lack of training data or the use of abstractive models is preferrable.
Outline of work plan: PHASE I (1st year):
- Step 1: study of the state-of-the-art Deep NLP embedding models (e.g., Word2Vec, GloVe, SIF, BERT, ElMo). Overview of state-of-the-art extractive summarization methods (e.g., SumPip, TextRankBM25, ELSA, Centroid) and sentiment analyzers (e.g., VADER).
- Step 2: retrieval of cross-lingual document corpora from various sources (e.g., Wikipedia). Analysis of the performance of state-of-the-art contextualized embedding models. Study, design, and developments of targeted model extensions to overcome specific issues (e.g., their limited applicability in cross-lingual and multi-domain scenarios).
Pursued objective: learn contextual embeddings on multilingual and multidomain data.

PHASE II (2nd year):
- Step 1: Study of new solutions for domain-adaptation and retrofitting of embedding models (i.e., tailor the models to specific domains). Empirical assessment on benchmark multilingual document corpora.
- Step 2: Application of the previously extended models to solve specific NLP tasks such as cross-lingual text summarization (extract a summary in a target language that differs from those of the source documents), cross-lingual sentiment analysis (exploit end-users opinions written in a given language to improve the quality of the sentiment analysis models specifically designed for other languages).
Pursued objective: develop new cross-lingual summarization and sentiment analysis methods.

PHASE III (3rd year):
- Step 1: Study and development of new abstractive summarization models (i.e., generate new text using Deep Learning architectures).
- Step 2: Study, design, and developments of state-of-the-art abstractive summarization methods. Evaluation on benchmark data.
Pursued objective: develop new abstractive summarization methods.

During all the three years the candidate will have the opportunity to attend top quality conferences, to collaborate with researchers from different countries, and to participate to NLP research challenges (e.g., FNS 2020 http://wp.lancs.ac.uk/cfie/fns2020/).
Expected target publications: Conferences: ACM SIGIR, EMNLP, COLING, IEEE ICDM, ACM KDD
Journals: IEEE TKDE, ACM TKDD, Elsevier ESWA, IEEE TETC, Springer DMKD
Current funded projects of the proposer related to the proposal: PRIN Italian Project Bando 2017 "Entrepreneurs As Scientists: When and How Start-ups Benefit from A Scientific Approach to Decision Making". The activity will be focused on the analysis of questionnaires, reviews, and technical reports related to training activities.
Possibly involved industries/companies:

Title: Analysis, search, development and reuse of smart contracts
Proposer: Valentina Gatteschi, Fabrizio Lamberti
Group website: http://grains.polito.it/
Summary of the proposal: In recent years, Blockchain technology received an increasing interest from researchers and companies. According to statistics, the global blockchain market is expected to grow from $5.8 billion in 2021 to more than $390 billion by 2028.
A key advantage of blockchain is that it can run smart contracts, (small) programs that are automatically executed (and that could trigger cryptocurrency transfers) when some conditions occur. Once a smart contract's code is stored on the blockchain, everyone can inspect it and it becomes no longer modifiable.
Despite the statistics, a considerable portion of the wider public still do not trust blockchain technology and smart contracts, mainly due to its complexity. In fact, understanding the behavior of a smart contract requires some (good) programming skills.
Another drawback of the current situation is that a comprehensive, easily-browsable repository to retrieve smart contracts’ code, in order to ease the development of new smart contracts is missing.
This proposal aims at addressing the above limitations by investigating new techniques and proposing novel approaches for the analysis, search, development and reuse of smart contracts. The results that will be achieved could be relevant for both developers, that could find/reuse smart contracts already developed in a given context, both people without technical skills, that could be able to understand the behavior of (or even code) a smart contract thanks to visualization and semantic technologies, among others. In addition to this main research topic, the Ph.D. candidate could address also other relevant blockchain-related issues such as scalability, security/privacy and interoperability.
Rsearch objectives and methods: The activities carried out in this Ph.D. programme will aim at investigating existing approaches, devising, and testing novel ones for:

a) the analysis of smart contracts’ source code/OPCODE: smart contracts’ source code/OPCODE deployed on the blockchain mainnet/testnet will be analyzed by considering different aspects, e.g., the quality of the written code, the behavior of the smart contract, the amount of triggered transactions, its context/objective, etc. The outcome of this activity will provide a methodology (or a tool) that could support developers and non-skilled people at a macro- and a micro-level. At a macro-level it could help them to understand how smart contracts are developed/used (or have been developed/used in the past), or to detect trends. At a micro-level, it could support them in quickly understanding the behavior of a smart contract. The analysis will be performed on the Ethereum blockchain, and eventually on other blockchains, to detect, for example, whether a given blockchain is preferred by a given sector, or to accomplish some specific tasks. Visual Analytics tools could be also exploited/developed to support the analysis.

b) the search of already developed smart contracts: a framework to categorize smart contracts analyzed in the previous phase will be designed. A methodology (or a tool) to search and retrieve the code of previously deployed smart contracts that could fulfill a given need will be devised. Semantics and Visual Analytics will be also used to support the search.

c) the development/reuse of smart contracts’ code: existing approaches for visual and natural language programming will be investigated to study their applicability to smart contracts. The result of this phase will be a methodology (or a tool) to support both programmers and non-skilled people in coding smart contracts from scratch or from existing ones (e.g., retrieved using the search engine devised in the previous phase).
Outline of work plan: The research work plan of the three-year Ph.D. programme is the following:
- First year: the candidate will perform an analysis of the state-of-the-art on available methodologies/tools for analysis, search and development/reuse of code, with a particular focus on smart contracts. The candidate will also deepen his/her competences on visual analytics, semantic technologies and natural language programming, among others. After the analysis, the candidate will identify advantages, disadvantages and limitations of existing solutions and will define approaches to overcome them.

- Second year: during the year, the candidate will design and develop methodologies and tools for analyzing smart contracts’ code, and will design a framework to categorize and search them.

- Third year: the third year will be devoted to the design and development of methodologies and tools enabling smart contracts reuse and coding, as well as to testing the developed tools.
Expected target publications:
Current funded projects of the proposer related to the proposal:
Possibly involved industries/companies:

Title: Dependable Computing in the heterogenous era: Fault Tolerance for the Middle Layer
Proposer: Alessandro Savino, Maurizio Rebaudengo
Group website: http://www.testgroup.polito.it
Summary of the proposal: Nowadays, the computing continuum is growing not only in computing capabilities but also in number of fields of application, placing Cyber-Physical Systems (CPSs) in a lot of heterogenous scenarios, ranging from agriculture to biology, including autonomous systems and IoT. The spreading comes with new challenges in which the dependability of the system is always crucial but the dependability constraints defined by each scenario vary a lot. For this reason, modern systems may require a middle layer that, by exploiting the highest level of fault tolerance, can enhance any application with a strong dependability. The proposed Ph.D. is going to focus on methodologies and tools to assess the dependability of Operating Systems (including Real-Time OSs), and to target the design of hardening techniques tailored to specific requirements imposed by each application field. The investigation will include exploiting all hardware features in modern microprocessors and systems to track the system’s behavior in order to support real-time adaptations.
Rsearch objectives and methods: The research objectives envisioned for this Ph.D. include:
1. To design and build fault injection strategies and tools to evaluate the OS resilience, resorting to all modern facilities, such as cloud-aware computation, architectural models, etc. All feasible solutions should be able to deliver detailed data about all single parts of the OS, including all system parameters evaluation, as well as to support highly scalable solutions to ensure the feasibility of the analysis.
2. To develop selective hardening techniques to empower the OS both of improved resiliency and higher filtering capabilities to reduce also the amount of errors rising at the application layer. The goal is to introduce techniques that, by being OS-aware and not generic, would improve the impact on power consumption and execution time with respect to the classical voting approaches on replicated executions.
3. To propose real-time online adaptation/reconfiguration techniques to dynamically react to non-masked faults and to allow preemptive behavior to future foreseeing issues.
The previous techniques are going to be investigated in a dynamic scenario, identifying fault detection techniques that, once triggered, induce the activation of countermeasures run-time, being able to carefully balance the impact of fault-tolerance techniques to other system parameters, such as power consumption and execution time.
Outline of work plan: Year 1: Dependability analysis of Hardware and Software Platforms

Reliability assessment is a crucial requirement to properly intervene within the OS and positively hardening the right parts. Thus, the Ph.D. student will likely spend the first year studying and designing solutions to assess the reliability of the system from an OS standing point. The aforementioned strategies will include the development of injectors (either based on architectural simulators, i.e., gem5, or pure software) able to properly inject all kind of faults, e.g., permanent, transient, and intermittent, to better investigate both the HW contribution to OS fault, and the OS masking capabilities. These tools will need to run on docker-like virtualization environments in order to allow an effective scalability of the fault injection campaigns

Year 2: Hardening the OS

Once detailed data are gathered, an in-deep analysis is capable to point out the unreliable components of the OS, paving the way to hardening those components and reducing the overheads introduced by the fault tolerant techniques. These techniques will be designed by using more selective and OS-aware solutions, like specific code protections, data reduction, and also replacing portion of the OS with intrinsic resilient techniques such as approximate computing, aiming at reducing the need of the classical duplication/triplication approaches. For these reasons, the expectations include models and tools to link the hardening solutions with all other parameters of the CPS, such as power consumptions, timing issues, and so on.

Year 3: Reconfigurable OS

Modern CPSs include many hardware facilities to online monitoring the system behavior. Those facilities may have predictive capabilities that, coupled with a reconfiguration capability of the OS, may deliver real-time adaption of the OS. The Ph.D. student will target the online monitoring and fault detection tasks and will develop reconfiguration methods to actively react on threads to the dependability detected at hardware level, also linkage their employment with power consumption and execution time. At this point, specific applications could targeted, representative of the most common embedded systems scenarios, to evaluate the flexibility of the run-time adaptation capabilities.
Expected target publications: We expect the candidate to be able to submit her/his work to several high-level conferences (e.g., DATE, DAC, DFT, ESWEEK, etc. ) as well as top ranking journals (e.g., IEEE Transaction on Computers, IEEE Transaction on Reliability). For the conference publication, we foreseen a minimum of two conference publications per year, while at least two publications, spawning along the second and the third year are expected.
Current funded projects of the proposer related to the proposal:
Possibly involved industries/companies:

Title: Beyond 5G (B5G) and the road to 6G Networks
Proposer: Claudio Casetti
Group website: http://www.dauin.polito.it/research/research_groups/netgroup_comp...
Summary of the proposal: 5G is expected to pave the way for the digitalisation and transformation of key industry sectors like transportation, logistics, commerce, production, health, smart cities, and public administration. While 5G has enabled us to consume digital media anywhere, anytime, the technology of the future should enable us to embed ourselves in entire virtual or digital worlds. In the world of 2030, human intelligence will be augmented by being tightly coupled and seamlessly intertwined with the network and digital technologies. B5G/6G shall aggregate multiple types of resources, including communication, data and AI processing that optimally connect at different scales, ranging from, e.g., in-body, intra- machine, indoor, data centres, to wide areas networks. Their integration results in an enormous digital ecosystem that grows more and more capable, intelligent, complex, and heterogeneous, and eventually creates a single network of networks. The PhD will focus on the investigation of the potentiality of B5G/6G communication and their impact on people’s lives, industrial development and societal advancement.
Rsearch objectives and methods: The PhD candidate will focus the research on one or more of the following directions:
- Novel zero-energy devices, with the challenge of making the network and devices smart enough to detect them without increasing EMF and energy consumption.
- Coexistence and cooperation of (non-3GPP and 3GPP) networks to be deployed through network densification using ubiquitous miniaturised, heterogeneous base stations with new MAC protocols for spatially multiplexed transmissions.
- Mechanisms and interfaces for intent-based direct wireless connectivity to be studied based on terminal trajectories or novel sensor-based intent detection for, e.g., Human-Machine Interfaces.
- Flexible resource planning by studying and characterizing the prediction of communication needs in combination with trajectory, resource and spectrum planning through system simulations of, e.g., Industry 4.0 settings.
- Resource planning and balancing of D2D and D2I resource assignments aided by AI algorithms.
Outline of work plan: The first six months (M0-M6) will be devoted to the establishment of the state-of-the-art in the field, looking at developments highlighted by the 6G Flagship consortium and to the narrowing down of the focus on one or more of the research objectives outlined above, placing them into the context of the 6G roadmap. The next six months (M7-M12) will therefore see the PhD candidate formulate a detailed research plan for the coming two years.

During the second year, until M24, the PhD candidate will work on the design and early development of tools (simulators, emulators, prototypes) for the study of the research objectives.

In the final year, until M36, the candidate will assess the work, review the objectives, quantify the Key Performance Indicators of the research and the benchmarks, leading to the writing of scientific papers and of the PhD thesis.
Expected target publications: Conferences:
IEEE Infocom, IEEE CCNC, ACM Mobicom, ACM Mobihoc

Journals:
IEEE Transactions on Mobile Computing
IEEE Transactions on Networks and Service Management
IEEE Communication Magazine
Current funded projects of the proposer related to the proposal: Hexa-X
Possibly involved industries/companies:

Title: Media Quality Optimization using Machine Learning on Large Scale Datasets
Proposer: Enrico Masala
Group website: http://media.polito.it
Summary of the proposal: An ever-increasing amount of data is available in digital format, a large part in the form of multimedia content, i.e., audio, images, and video. The perceptual quality of such data varies widely, depending on the type of content (professional or user-generated), equipment used, bandwidth, storage space, etc. However, users’ quality of experience when dealing with multimedia content strongly depends on the perceived quality, therefore a number of algorithms and techniques have been proposed in the past for quality prediction. Recently, machine learning significantly contributed to the effort (e.g., Netflix VMAF proposal). However, despite the improvements, no technique can currently be considered really reliable, partly because the inner workings of machine learning (ML) models cannot be easily understood.
The activity proposed here will focus on optimizing the media compression and communication scenario. Particular attention will be devoted to the key objectives of this proposal, i.e., the creation of tools for a systematic, large-scale approach to the problem and the identification of media characteristics and features that can explain the typical black box ML models. The final aim is to find the characteristics that most influence perceptual quality in order to improve existing measures and algorithms on the basis of such new knowledge.
Rsearch objectives and methods: A few ML-based techniques have been recently proposed for media quality optimization, however almost all of them work as a black box without giving hints about what could be changed to improve performance.
Starting from our ongoing work on modeling subjective media quality experiments in terms of the behavior of single human subjects, a systematic approach will be followed by analyzing and taking advantage of the several subjectively annotated databases available in literature. The final aim is to obtain improved and explainable media quality features that can be combined into reliable quality predictions.
Moreover, large-scale unannotated datasets will also be considered. To this aim, we believe it is necessary to develop a framework comprising a set of tools that allows to more easily process subjective scores (given by human subjects) as well as objective ones in an efficient and integrated manner. Such framework, that we plan to make publicly available for research purposes, will constitute the basis for reproducible research, which is increasingly important for ML techniques.
The framework will allow to systematically investigate existing quality prediction algorithms finding strength and weaknesses, as well as to identify the most challenging content on which newer development can be based.
The large set of data obtained by such a systematic investigation is expected to facilitate the identification of a set of features that can be considered as candidates to explain the reason of subjective quality scores.
Such objectives will be achieved by using both theoretical and practical approaches. The resulting insight will then be validated in practical cases by analyzing the performance of the system with simulations and experiments with industry-grade signals, leveraging the ongoing cooperation with companies to facilitate the migration of the developed algorithms and technologies into prototypes that can then be effectively tested in real industrial media processing pipelines.
Outline of work plan: In the first year, the PhD candidate will first familiarize with the recently proposed ML and AI-based techniques for media quality optimization, as well as the characteristics of the publicly available datasets for research purposes.
In parallel, a framework will be created to efficiently process the large sets of data (especially for the video case) with potentially complex measures that might need retraining, fine-tuning or other computationally complex optimizations. It is expected to make this framework publicly available also to address the research reproducibility issues that are of growing interest in the ML community. This initial investigation and activities are expected to lead to conference publications.

In the second year, building on the framework and the theoretical knowledge already present in the research group, new media quality indicators for specific quality features will be developed, simulated, and tested to demonstrate their performance and in particular their ability to identify the root causes of the quality scores for several existing quality prediction algorithms, thus partly explaining their inner working methods in a more understandable form. In this context, potential shortcomings of such algorithms will be systematically identified. These results are expected to yield one or more journal publications.

In the third year the activity will then be expanded to propose improvements that can mitigate the identified shortcoming as well as to create proposals for quality prediction algorithms based on the previously identified robust features. Such proposal will target journal publications.
Expected target publications: Possible targets for research publications (well known to the proposer) include IEEE Transactions on Multimedia, Elsevier Signal Processing: Image Communication, ACM Transactions on Multimedia Computing Communications and Applications, Elsevier Multimedia Tools and Applications, various IEEE/ACM international conferences (IEEE ICME, MMSP, QoMEX, ACM MM, ACM MMSys).
Current funded projects of the proposer related to the proposal: PIC4SeR Interdepartmental Center for Service Robotics (1 PhD scholarship - XXXV cycle - already in place), Sky Group (collaboration agreement)
Possibly involved industries/companies:

Title: Digital Wellbeing in the Internet of Things
Proposer: Luigi De Russis
Group website: https://elite.polito.it
Summary of the proposal: While people derive several benefits from using a plethora of “smart” devices every day, the last few years have seen a growing amount of attention on the negative aspects of overusing technology.
This attention has resulted in the so-called “digital wellbeing”, a term that refers to the impact of digital technologies on people’s lives, i.e., “what it means to live a life that is good for a human being in an information society”.

Models and tools related to the digital wellbeing context, however, often considers single technological sources at a time, mainly the smartphone. Targeting a single source is not sufficient to capture all the nuances of people’s digital wellbeing: in today’s multi-device world, indeed, people use more than one device at a time, and more effort should be put into evaluating multi-device and cross-device interaction to enhance digital wellbeing.

This PhD proposal investigates digital wellbeing in the Internet of Things (IoT), with the aim of providing insights, strategies, tools, and interfaces to better cope with digital wellbeing in our multi-device and connected environments.
Rsearch objectives and methods: The main research objective is to explore novel and effective ways to cope with digital wellbeing in IoT ecosystems, where people may use multiple devices at a time for different purposes. The PhD student, in particular, will work on the study, design, development, and evaluation of proper models and novel technical solutions for this domain (e.g., tools and frameworks), starting from the relevant scientific literature and performing studies directly involving users in multi-device settings.

Possible areas of investigation for digital wellbeing solutions include:
a) Data integration. Tools should be able to make sense of data coming from different devices to provide users with a high-level overview of their technological habits. The variety of IoT devices we have today, in particular, opens up new possibilities: a tool could exploit cameras and smartphone sensors in a sensible and privacy-preserving way to enrich usage data with contextual information, e.g., to figure out where a person is looking at. Understanding what the user is currently doing with their devices would be extremely important, as the underlying task is a discriminant factors to differentiate between positive and negative multi-device experiences.
b) Cross-device interventions. Tools should adapt to the characteristics of the target device, to allow users to control different device-specific behaviors, from managing smartphone notifications to avoiding excessive zapping on the smart TV. Data integration could allow the activation of interventions on the basis of the current activity, e.g., to block smartphone distraction while working at the PC.
c) Learning. Contemporary digital wellbeing solutions primarily focus on lock-out mechanisms to reduce device use. While those mechanisms are the shortest path to avoid unwanted behaviors, they also proved not to be effective. Instead of blocking “bad” behaviors, a novel solution could be used as a learning support, e.g., by suggesting desirable alternatives to scaffold new habits.
Outline of work plan: The work plan will be organized according to the following four phases, partially overlapped.
Phase 1 (months 0-6): literature review about digital wellbeing in various contexts (e.g., home, work, …); study and knowledge of IoT devices and smart appliances, as well as related communication and programming practices and standards.
Phase 2 (months 6-18): based on the results of the previous phase, definitions and development of a set of use cases, interesting contexts, and promising strategies to be adopted. Initial data collection for validating the identified strategies in some contexts and use cases.
Phase 3 (months 12-24): research, definition, and experimentation of multi-device technical solutions for digital wellbeing, starting from the outcome of the previous phase. Such solutions will likely imply the design, implementation, and evaluation of distributed and intelligent systems, able to take into account both users’ preferences and capabilities of a set of connected devices.
Phase 4 (months 24-36): extension and possible generalization of the previous phase to include additional contexts and use cases. Evaluation in real settings over long period of times to assess at which extent the proposed solutions are able to address the negative impact and increase any positive outcome on our digital wellbeing.
Expected target publications: It is expected that the results of this research will be published in some of the top conferences in the Human-Computer Interaction field (ACM CHI, ACM CSCW, ACM Ubicomp). One or more journal publications are expected on a subset of the following international journals: ACM Transactions on Computer-Human Interaction, ACM Transactions on Cyber-Physical Systems, International Journal of Human Computer Studies, IEEE Internet of Things Journal, ACM Transactions on Internet of Things, IEEE Transactions on Human-Machine System.
Current funded projects of the proposer related to the proposal: None at the moment
Possibly involved industries/companies:None at the moment

Title: Advancing Mobile UI Testing Techniques
Proposer: Marco Torchiano, Luca Ardito
Group website: http://softeng.polito.it/
Summary of the proposal: Testing of mobile application user interfaces (UI) consists in writing test cases that exercise the UI and allow performing end-to-end testing of the whole app in an automated way. Tests for mobile applications are particularly challenging since they have to address a wide range of devices and platforms.
Despite the presence of several mature frameworks, UI tests show a limited adoption. This is certainly due to the significant effort in writing the tests cases, though the main barrier is represented by the inherent fragility of such kind of tests: a minor change in the UI can “break” the test requiring a significant effort in identifying the cause and then fixing the test.
Lowering the adoption barrier requires both working on existing test suites with the ability to detect fragile tests in existing suites and then techniques to fix such fragilities; in addition novel approaches to build tests ought be devised that minimize writing effort, and enable effective strategies to evolve test suites.
Rsearch objectives and methods: O1.Identification of test fragilities
This step is about the definition of techniques to detect the patterns that cause test fragility. A comprehensive taxonomy (O1.1) is the prerequisite. The automatic detection of fragilities (O1.2) can then be developed. A tool that can work as a plug-in of an IDE represents the main outcome.

O2.Test fragility mitigation
Building upon the results of O1, we need to identify remediation techniques for the fragility-inducing patterns (O2.1). The remediation techniques must be implemented in order to achieve automatic removal (O2.2). The implementation should find its place within a plug-in for an IDE.

O3.Definition of novel testing techniques
The identification of the fragility-inducing patterns represents the basis also for a gap analysis of existing techniques (O3.1). Novel techniques should be defined (O3.2) typically by leveraging the relative strengths of existing ones and using novel techniques, e.g. machine vision to generate more accurate tests.

O4.Definition of test evolution strategies
The key limitation of UI tests is represented by the need for evolving along with the application UI. A set of recommendations and strategies must be devised to guide the evolution of both existing and novel tests. In this context Machine Learning techniques can be used to provide precise recommendations for test cases adaptation.
Outline of work plan: The main activities conducted in the three years of the PhD studies are:
Task 1.1: Development of a fragility detection tool
Task 1.2: Empirical assessment of effectiveness of fragility detection
Task 2.1: Identification of the fragility mitigation techniques
Task 2.2: Development of a fragility removal tool
Task 2.3: Empirical assessment of fragility removal
Task 3.1: Analysis of limitation of existing tools and techniques
Task 3.2: Definition of a novel (integrated) testing approach
Task 4.1: Identification of a test evolution strategy
Task 4.2: Empirical assessment of test evolution strategy


The four work units will follow similar methodological approaches where
- an initial identification of the issues is conducted by means of observational methods (case studies, surveys, case control studies);
- a solution is proposed
- a proof of concept implementation is developed
- an empirical assessment is conducted typically using controlled experiments.
Expected target publications: The target for the PhD research includes a set of conferences in the general area of software engineering (ICSE, ESEM, EASE, ASE, ICSME) as well as in the specific area of testing (ICST, ISSTA).
More mature results will be published in software engineering journal, the main being: IEEE Transactions on Software Engineering, ACM TOSEM, Empirical Software Engineering, Journal of Systems and Software, Information and Software Technologies.
Current funded projects of the proposer related to the proposal:
Possibly involved industries/companies:Several companies expressed concerns on mobile UI tests, among them three showed more interest in the specific goals of this PhD proposal: Reale Mutua Assicurazioni, Intesa SanPaolo, Engineering

Title: Cybersecurity applied to embedded control systems
Proposer: Stefano Di Carlo and Alessandro Savino
Group website: https://www.testgroup.polito.it
Summary of the proposal: Most of today’s embedded systems include microprocessor cores able to efficiently perform an always increasing number of complex tasks. These electronic devices usually include some microprocessor cores, different memory cores, and in some cases custom logic blocks able to support the final application requirements. Manufacturer companies of embedded systems have been asked to speed up the time to market while increasing microprocessor speed, reducing power consumption, improving system security, and reducing production costs.
One of the most important issues faced by the manufacturers of embedded systems is related to the appropriate methodologies to guarantee safe and secure devices; indeed, new security issues are required to be considered in the early stages of the processor design cycle.
During this project, the Ph.D. candidate will study and provide solutions regarding how to improve the security aspects of embedded systems.
Rsearch objectives and methods: Embedded computing systems (ECS) must be designed having security, privacy, data protection, fault tolerance and accountability in mind.

In particular, the hardware is the root of trust of a full ECS; if the hardware is jeopardized, the whole system is at stake. Thus, it is imperative to enhance control and trust of hardware across the supply chain and during its operation, and this is the global objective of this Ph.D. project.

More in details, this Ph.D. project aims at investigating a new CPU-centric architecture protected against the major hardware security threats observed across different computing domains: from low-end embedded to High Performance Computing (HPC) systems, with particular emphasis on systems employed in the automotive domain.

The main result pursued by the research activities is the development of a holistic analysis framework and a complete system-level implementation strategy to guarantee security throughout the hardware design/manufacturing supply chain and the system operational lifetime through the integration of dedicated hardware and software blocks able to remove or minimize the impact of the Integrated Circuits (IC) supply chain vulnerabilities as well as possible physical and functional side-channel attacks in the field.

The combined/unified investigation of the different vulnerabilities passes through a detailed analysis and modeling of the generic system-level “perturbations” or “disturbances” occurring to the normal system operation when it is being attacked either by malicious pre-designed hardware structures or through physical or functional attacks during system operation. This in turns requires to create a unified taxonomy of the pre- and post-deployment security threats. Based on this taxonomy, the identified hardware security vulnerabilities will be prioritized, and corresponding hardware will be developed and evaluated.

The project aspires to work on open hardware architectures such as RISC-V but also on proprietary architecture in collaboration with major players in selected application domains (e.g., automotive domain).
Outline of work plan: The proposed research activities are structured over a three years research program.

1st year: Toolchain, Taxonomy and Characterization.

The first year of activity focuses on the development of the infrastructure required to develop, analyze, evaluate and demonstrate the idea of the security custodians through the supply chain of CPU-based ECS. This includes both: (a) the physical nodes (selected hardware cores and supporting software modules) on which all protection mechanisms will be developed and analyzed, and (b) a flexible simulation/emulation platform for porting and trade-off analysis. The two pieces (actual hardware and simulation/emulation) will complement each other throughout the project duration. The impact of the hardware custodians in terms of performance, power, energy, and reliability will be evaluated by comparing several well-established benchmarks such as the SPEC, PARSEC, EEMB SecureMark, or Apollo on the original and protected chips and the ability to counteract the considered attacks will be evaluated by means of well targeted applications identified on an attack base (e.g., cryptographic algorithms). Together with this activity, the first year of the Ph.D. will be devoted to the compilation of a complete taxonomy of hardware vulnerabilities through the whole ECS supply chain. Vulnerabilities will be prioritized based on their severity in different application domain in order to drive the activities of the second year.

2st year: Design for security.

The second year of the research program is dedicated to the development of specific hardware and software custodians to counteract the most relevant hardware vulnerabilities identified during the first year. Several classes of attacks are considered including information leaking and denial of service attacks performed by exploiting hardware vulnerabilities. Particular attention will be dedicated to those attacks impacting the memory hierarchy of the system that recently gained significant attention.

Every developed custodian will be analyzed resorting to the simulation/emulation platform developed during the first year or through real prototype hardware platforms (e.g., FPGAs). The analysis will not be limited to the provided security but also extended to other design dimensions (e.g., performance, power, reliability, etc.).

The main outcome of the research activity of the second year will be the development of a library of hardware and software custodians that can be plugged in an ECS hardware design to protect against selected attacks.

3st year: Integration and demonstration.

The last year of the Ph.D. will mainly focus on integrating all solutions developed during the first two years in order to develop a complete design for security flow based on the hardware custodians.

To demonstrate the effectiveness of the developed design flow, the full architecture will be demonstrated considering a real use case. Among the different application scenarios, we will focus on embedded control systems for the automotive domain.
Expected target publications: All activities carried out during the three years research program will be the subject of scientific publication.
In particular the publication activities will start with the submission of a complete review paper on the taxonomy of relevant hardware vulnerabilities for ECS to a high-quality journal.

In parallel, results obtained from the development of the analysis framework, implementation and evaluation of the security custodians will be subject of publications at major conferences to obtain fast feedbacks from the scientific community regarding the proposed solutions.

The main conferences where the Ph.D. student will possibly publish her/his works are:
DATE: IEEE - Design, Automation and Test in Europe Conference
DAC: IEEE Design Automation Conference
ETS: IEEE European Test Symposium
ITC: IEEE International Test Conference
VTS: IEEE VLSI Test Symposium
IOLTS: IEEE International Symposium on On-Line Testing and Robust System Design
MTV: IEEE International workshop on Microprocessor/SoC Test, Security and Verification

Eventually, different aspects of the complete framework will be subject of publications submitted to high-quality archival journals such as IEEE TCAD, IEEE TVLSI, IEEE ToC.
Current funded projects of the proposer related to the proposal:
Possibly involved industries/companies:Punch Torino

Title: Planning Safety in the Era of Approximate Computing Systems
Proposer: Stefano Di Carlo, Alessandro Savino
Group website: http://www.testgroup.polito.it
Summary of the proposal: Cyber-Physical Systems (CPSs) are the root of the fourth industrial revolution and the Internet of Things (IoT) forms a foundation for this cyber-physical systems revolution. CPSs are large physical systems that consist of many interacting physical elements and distributed IT systems for monitoring, control, optimization and interaction with human operators and managers. All devices of a CPS and different CPSs are connected in order to create a network enabling billions of systems and devices to interact and share information.
While CPSs are crucial to many application domains, several modern computations (such as video and audio encoders, Monte Carlo simulations, and machine learning algorithms) are designed to trade off accuracy in return for increased performance and lower power consumption. To date, such computations typically use ad-hoc, domain-specific techniques developed specifically for the computation at hand and none of them consider the final safety impact on the CPSs.
In this context, this Research Proposal is particularly interested at analyzing the safety of CPSs when approximate computing is used. The main goal is to analyze several approximation paradigms when dealing with safety constraints and to propose alternative strategies to cope with any safety issue.
Rsearch objectives and methods: The research objectives envisioned for this work include both approximate computing and safety aspects:

1. To develop approximate computing methods to assess the impact on safety. Those methods will properly guide the application of approximation under safety constraints. The Safety standards should come from the most interesting fields of application, such as automotive and aerospace.
2. To study and to propose Safety-oriented approximate computing models to target an early analysis of the system under approximation usage. Those models must enhance the capability of early designing approximated CPS under safety constrains by playing with all system parameters. The parameters should include all layers of a CPS, such as hardware components and their specific instantiation (e.g., CPUs, GPUs, FPGA, etc.) as well as middleware (e.g., Real-Time Operating systems) and software.
3. To provide tools to carefully design CPS including the approximate computing without failing safety constraints. Such tools should include design space exploration techniques to balance the larger number of design constraints, such as power consumption, area, time, etc. Previously developed methods and models should empower those tools with the necessary flexibility in meeting the constrains and system final design.

Those objectives could also lead to the participation or proposition of new Safety standards.
Outline of work plan: Year 1: Safety standards and approximate computing estimation methods

The first step to achieve the goals described in the previous section is to analyze all current approximate computing strategies alongside with the current Safety standards to develop effective estimation methods to properly link the two worlds.

To carry out this activity the student requires to acquire a deep knowledge of the state-of-the-art in order to identify safety-related case-studies and Approximate Computing techniques to be applied. The estimation methods should evaluate the accuracy and the safety as well as the power, performance.

Year 2: From Methods to Models and Paradigms

The second step carries out the transition from the estimation methods to the proper modeling of the CPS system in order to ease the analysis of the safety impact. The student must become familiar with several modeling, programming and simulation frameworks at different layers. Artificial Intelligence models and Machine Learning techniques could be used to properly analyze the models, and the definition of new paradigms for the evaluation can be part of this step of the work.

Year 3: Tools, Design Space Exploration and further model improvements

The last year of the Ph.D. will be dedicated to the development of tools and Design Space Exploration (DSE) techniques for optimizing CFPs under safety constrains. These approaches will exploit the capability of the models developed during the second year of the Ph.D. project coupled with efficient space exploration algorithms for multi-objective optimization such as the extremal optimization theory or other evolutive approaches.

The DSE will be fed with multi-objective goals in order to provide an instrument to trade-off safety and accuracy with power and performance.
Expected target publications: The work developed within this project can be submitted to:

Conferences (DSN, DATE, DAC, IOLTS, etc.): a minimum of two conference publications per year.

Journals: IEEE TOC, IEEE TCAD, ACM TACO, IEEE TDSC: at least one publication during the second year and one during the third year.
Current funded projects of the proposer related to the proposal:
Possibly involved industries/companies:Intel Corporate Italia

Title: Exploring Deep Learning for In-Field Fault Detection in Cyber Physical Systems
Proposer: Alessandro Savino, Stefano Di Carlo
Group website: http://www.testgroup.polito.it
Summary of the proposal: Rapid innovation in strategic industrial fields such as medical, automotive, IoT, and HPC, is pushing for the adoption of high-scaled multi-core processors in several domains. This puts pressure on systems’ designers and manufacturers to deliver improved availability and reliability starting from the early design phases. Recent advances in Deep Learning and Artificial Intelligence may provide new powerful tools to build advanced fault detection systems supported by the development of fault injection frameworks able to collect huge amount of simulated data.
The proposed Ph.D. project will explore the feasibility of using deep learning models for the design of a monitoring mechanism able to detect permanent and transient faults in microprocessors executing software applications. Particular emphasis will be put on domain adaptation capabilities and time series analysis.
Rsearch objectives and methods: The main contribution of this proposal is investigating how deep learning models can be exploited to detect permanent and transient faults in microprocessors executing software applications. In particular, candidates are expected to study how these methods are able to (i) generalize over several domains represented here by different software applications running on the same hardware, thus minimizing retraining, and (ii) work with a set of low-level features that can be collected at the microprocessor architectural and microarchitectural level, allowing hardware implementations of the trained models.

The research objectives envisioned for this Ph.D. include:
1. To develop an emulation framework able to collect microarchitectural feature from the execution of software applications on multi-core microprocessors.
2. To develop design and investigation methodologies targeting Deep Learning techniques for the In-Field Fault Detection of hardware faults, including the support for generalization over several domains and the evaluation of time series data.
3. To define new microprocessor infrastructures for efficient monitoring of relevant features.
4. To propose methods and tools to embed the detection capabilities at the hardware and operating system level.
Outline of work plan: Year 1: Deep Learning Methodologies and Data Generation Tools

The first step requires investigating how Deep Learning models are able to generalize over microarchitecture level features collected from the execution of different applications on the same microprocessor, thus minimizing retraining. Depending on the characteristics of the assembly code of the selected applications, similar activation patterns can be generated for the different functional units of the microprocessor thus facilitating the domain adaptation.

To achieve this goal, a fully automated flow is necessary, able to collect data, to train and test the detection model and to validate detection prediction. The main challenge here is to be able to emulate the occurrence of selected classes of hardware faults and to observe selected microprocessor features. This requires building a full emulation framework allowing high-throughput fault injection experiments that can be used to collect data required to build machine learning models.

Year 2: Full Detection Support

Once useful models are identified, in order to fully exploit their detection capability, all required features should be collectable at run-time. Nowadays, commercial processors are not ready for this activity, and only limited run-time information is collected by means of Performance Monitoring Counters (PMC).

The main activity of the second year is to design dedicated hardware infrastructure working on open microprocessor architectures (e.g., RISC V), able to trace and collect relevant features identified during the first year and not already accessible in the considered architectures.

Year 3: Real-Time monitoring

Once a hardware infrastructure to profile the execution features is available, an in-field monitor implementing the identified deep learning models can be embedded in the system. This includes adapting available operating systems to properly access the monitoring infrastructure defined during the second year, and implementing the identified learning model in such a way to implement a run-time fault detection system. Dedicated hardware co-processor able to unload the CPU from the burden of implementing this detection task will also be investigated.
Expected target publications: We expect the candidate to be able to submit her/his work to several high-level conferences (e.g., DATE, DAC, DFT, ESWEEK, etc. ) as well as top ranking journals (e.g., IEEE Transaction on Computers, IEEE Transaction on Emerging Topics in Computing).
Current funded projects of the proposer related to the proposal:
Possibly involved industries/companies:

Title: Testing and Reliability of Automotive oriented System-on-Chip
Proposer: Paolo Bernardi, Riccardo Cantoro
Group website: cad.polito.it
Summary of the proposal: Nowadays, the number of integrated circuits included in critical environments such as the automotive field is continuously growing. For this reason, semiconductor manufacturer has to guarantee the reliability of the released components for the entire life-cycle that can be up to 10 -15 years.

The activities planned for this proposal includes efforts towards
- Study and development of innovative Fault Simulation strategies that could manage current designs complexity
- The optimization of crucial Reliability Measurements such as Burn In, System Level Test and Accelerated Operating Life Tests, in addition to circuit Repair strategies
- The design of test strategies aimed at supporting in-field error detection which are demanded by recent standards such as the ISO 26262
- The application of Machine Learning methodologies to elaborate diagnostic manufacturing volume result
- The analysis of the Reliability issues of Accelerator Cores implementing Artificial Intelligence on-chip.

The project will enable a phd student to work on a very timely subject and supported by companies currently collaborating with the research group. Expectation is to actively contribute to next generation chip reliable design.

The potential outcome of the project is both related to industrial advances and high quality publication.
Rsearch objectives and methods: The phd student will pursue the following objectives in the research activities. All of them are part of the broader scenario of the Automotive Reliability and Testing field; even if this scenario is varying in time, the current objective are looking to the most significant challenges.

Reliability measurement
- Improve the effectiveness of TEST and STRESS pattern generation through thermal/power analysis and modeling
1. Better coverage of defects such as delay and exacerbation of intermittent faults
2. Reduction of the Burn-In time
3. Better effectiveness of System-Level Test procedures
- Introduction of innovative flows to increase the yield in the manufacturing process
- Novel repair algorithms for embedded cores (i.e., memory cores)
- Setup a low-cost test scenario based on low-cost equipment and computational infrastructure.
1. Design and implementation of a tester able to effectively drive the test and diagnosis procedure
2. Logic diagnosis based on the collected results on a set of failing devices
3. Provide information to Failure Analysis labs for a faster identification of the root cause of a malfunctioning

In-field testing
- Development of methods for achieving high coverage of defects appearing along mission behavior as demanded by the ISO-26262
1. Key-on and runtime execution of Software-based Self-Test procedure for CPUs and peripheral cores
2. Diagnosis trace for faster failure analysis of returns from fields.

Machine learning methodologies applied to test
- Conception and implementation of machine learning methodologies to elaborate diagnostic
manufacturing volume result
- Reconstruction of failure bitmaps after a compression phase
- Prediction of the stress level through indirect measurement on chip

Reliability issues of Accelerator Cores
- HW Accelerators are becoming more and more important and their usage in safety critical field is raising questions about their reliability
- SW libraries development for such kind of embedded cores will be investigated
Outline of work plan: The working plan for the PhD student is recalling the objectives drawn in the previous sections. The order is not fixed and may vary according to the advancement during the PhD program.

1st year
1. SoC sensitivity evaluation, mission mode FIT measurement and estimation through fast methods
2. Low-Cost Tester for Delay and Transient faults, based on multi-processor cores
3. Repair strategies overview for SoC components, (especially for embedded memories)
4. Volume data analysis and implementation of ML to memory testing
5. Analysis of techniques used for guaranteeing AI hardware modules reliability

2nd year
6. Diagnosis strategies development for quickly feedback failures appearing along Burn In or causing field returns
7. Reconstruction techniques for uncomplete failure bitmaps
8. Logic diagnosis of failing devices, with the intention of providing a fault location and fault model hypothesis to failure analysts
9. Design of software libraries for AI modules

3rd year
10. Advanced techniques for design and implementation for reliable and safe Automotive ICs. Design of a SW platform that elaborated the failure information wrote in the DB
Expected target publications: IEEE Transactions on Computers, IEEE Transactions on Emerging Topics in Computing, IEEE Design and Test, IEEE Access
Current funded projects of the proposer related to the proposal: In this topic, the proposer is involved in research programs funded by STMicroelectronics, Infineon
Possibly involved industries/companies:STMicroelectronics, Infineon, Xilinx

Title: Graph network models for Data Science
Proposer: Elena Baralis, Daniele Apiletti
Group website: https://dbdmg.polito.it
Summary of the proposal: Machine learning approaches extract information from data with generalized optimization methods. However, besides the knowledge brought by the data, extra a-priori knowledge of the modeled phenomena is often available. Hence an inductive bias can be introduced from domain knowledge and physical constraints, as proposed by the emerging field of Theory-Guided Data Science.
Within this broad field, the candidate will explore solutions exploiting the relational structure among data, represented by means of Graph Network approaches.
Relational structure is present in many real-life settings, both in physical conditions, such as among actors in supply chains or users in social networks, and logical processes performed by humans, such as industrial procedures.
The structure of the data can be exploited to directly build the network graph itself, incorporating hierarchies and relationships among the different elements.
Analogous approaches can be exploited for logical processes where domain experts separate the overall procedure in connected subtasks for their decision making.
Hence, a graph-like structure can be crafted to design an ensemble architecture consisting of different building-blocks, each connecting a network node representing the sub-problem and the corresponding domain-driven knowledge and constraints.
The candidate should explore such a set of approaches to design and evaluate innovative learning strategies able to blend domain-expert behaviors, a-priori knowledge, and physical or theoretical constraints with the traditional data-driven training.
Rsearch objectives and methods: The aim of the research is to define new methodologies for semantics embedding, propose novel algorithms and data structures, explore applications, investigate limitations, and advance the solutions based on different emerging Theory-guided Data Science approaches.
The final goal is to contribute to improving the Machine Learning model performance by reducing the learning space thanks to the exploitation of existing domain knowledge in addition to the (often limited) available training data, pushing towards more unsupervised and semantically richer models.
To this aim, the main research objective is to exploit the Graph Network frameworks in deep-learning architectures by addressing the following issues:
- Improving state-of-the-art strategies of organizing and extracting information from structured data.
- Overcoming the Graph-Network model limitation in training very deep architectures, with a consequent loss in expressive power of the solutions.
- Advancing the state-of-the-art solutions to dynamic graphs, which can change nodes and mutual connections over time. Dynamic Networks can successfully learn the behavior of evolving systems.
- Experimentally evaluate the novel techniques in large-scale systems, such as supply chains, social networks, collaborative smart-working platforms, etc. Currently, for most graph-embedding algorithms, the scalability of the structure is difficult to handle since each node has a peculiar neighborhood organization.
- Applying the proposed algorithms to natively graph-unstructured data, such as texts, images, audio, etc.
- Developing techniques to design ensemble graph architectures to capture domain-knowledge relationships and physical constraints.
Outline of work plan: 1st year. The candidate will explore the state-of-the art techniques of dealing with both structured and unstructured data, to integrate domain-knowledge strategies in network model architectures. Applications to physics phenomena, images and text, taken from real-world networks such as social platforms and supply chains will be considered.

2nd year. The candidate will define innovative solutions to overcome the limitations described in the research objectives, by experimenting the proposed techniques on the identified real-world problems. The development and the experimental phase will be conducted on public, synthetic, and possibly real-world datasets. New challenges and limitations are expected to be identified in this phase.

During the 3rd year the candidate will extend the research by widening the experimental evaluation to more complex phenomena able to better leverage the domain-knowledge provided by the Graph Networks. The candidate will perform optimizations on the designed algorithms, establishing limitations of the developed solutions and possible improvements in new application fields.
Expected target publications: IEEE TKDE (Trans. on Knowledge and Data Engineering)
ACM TKDD (Trans. on Knowledge Discovery in Data)
ACM TOIS (Trans. on Information Systems)
ACM TOIT (Trans. on Internet Technology)
ACM TIST (Trans. on Intelligent Systems and Technology)
IEEE TPAMI (Trans. on Pattern Analysis and Machine Intelligence)
Information sciences (Elsevier)
Expert systems with Applications (Elsevier)
Engineering Applications of Artificial Intelligence (Elsevier)
Journal of Big Data (Springer)
ACM Transactions on Spatial Algorithms and Systems (TSAS)
IEEE Transactions on Big Data (TBD)
Big Data Research
IEEE Transactions on Emerging Topics in Computing (TETC)
Information sciences (Elsevier)
Current funded projects of the proposer related to the proposal: Research contract “Data Science and Machine Learning techniques for clinical supply chains”.
Possibly involved industries/companies:XelionTech, FBK, Istituto Mario Negri.

Title: Promoting Diversity in Evolutionary Algorithms
Proposer: Giovanni Squillero
Group website: http://www.cad.polito.it/
Summary of the proposal: Computational intelligence (CI) in general, and evolutionary computation (EC) in particular, is experiencing a peculiar moment. On the one hand, fewer and fewer scientific papers focus on EC as their main topic; on the other hand, traditional EC techniques are routinely exploited in practical activities that are filed under different labels. Divergence of character, or, more precisely, the lack of it, is widely recognized as the most impairing single problem in the field of EC. While divergence of character is a cornerstone of natural evolution, in EC all candidate solutions eventually crowd the very same areas in the search space, such a “lack of speciation” has been pointed out in the seminal work of Holland back in 1975. It is usually labeled with the oxymoron “premature convergence” to stress the tendency of an algorithm to convergence toward a point where it was not supposed to converge to in the first place. The research activity would tackle “diversity promotion”, that is either “increasing” or “preserving” diversity in an EC population, both from a practical and theoretical point of view. It will also include the related problems of defining and measuring diversity.
Rsearch objectives and methods: This research started about a decade ago with “A novel methodology for diversity preservation in evolutionary algorithms” (G. Squillero, A. Tonda; GECCO Companion; 2008) and is more active every passing year. The proposer organized workshops on diversity promotion in 2016 and 2017; moreover, he presented tutorials on the subject at IEEE CEC 2014, PPSN 2016, and ACM GECCO 2018. This year, together with Dirk Sudholt, a scholar devoted to theoretical aspects of EAs, the research incorporated rigorous runtime analyses. A joint tutorial was presented at ACM GECCO 2020 (“Theory and Practice of Population Diversity in Evolutionary Computation”).

The primary research objective will be to choose a possible definition of diversity, and to analyze and develop well-known, highly-effective, general-purpose methodologies able to promote it. That is, methodologies like the “island-model”, that are not linked to a specific implementation nor to a specific paradigm but are able to modify the whole evolutionary process. Then the study would examine why and how the divergence of character works in nature, and then find analogies in the artificial environment. Such ideas would be first tested on reduced environment and problems used by scholars. Later, they will be generalized to a wide scenario, broadening their applicability for practitioners. Rigorous runtime analyses would be used to assess the proposed methodologies.

As “premature convergence” is probably the single most impairing problem in the industrial application of EC, any methodology able to ease it would have a tremendous impact. To this end, the proposed line of research is generic and deliberately un-focused, not to limit the applicability of the solutions. However, the research will explicitly consider domains where the proposer has some experience. Namely:
- CAD Applications, mostly related to the generation of Turing-complete assembly programs for test and validation of microprocessors.
- Evolutionary Machine Learning, that is mostly EC techniques used to complement traditional ML approaches.
- Computational Intelligence and games
Outline of work plan: The first phase of the project shall consist of an extensive experimental study of existing diversity preservation methods across various global optimization problems. The MicroGP, a general-purpose EA, will be used to study the influence of various methodologies and modifications on the population dynamics. Solutions that do not require the analysis of the internal structure of the individual (e.g., Cellular EAs, Deterministic Crowding, Hierarchical Fair Competition, Island Models, or Segregation) shall be considered. This study should allow the development of a, possibly new, effective methodology, able to generalize and coalesce most of the cited techniques.

During the first year, the candidate will take a course in Artificial Intelligence, and all Ph.D. courses of the educational path on Data Science. Additionally, the candidate is required to learn the Python language.

Starting from the second year, the research activity shall include Turing-complete program generation. The candidate will move to MicroGP v4, the new, Python version of the toolkit under active development. That would also ease the comparison with existing state-of-the-art toolkits, such as inspyred and deap. The candidate will try to replicate the work of the first year on much more difficult genotype-level methodologies, such as Clearing, Diversifiers, Fitness Sharing, Restricted Tournament Selection, Sequential Niching, Standard Crowding, Tarpeian Method, and Two-level Diversity Selection.

At some point, probably toward the end of the second year, the new methodologies will be integrated into the Grammatical Evolution framework developed at the Machine Learning Lab of University of Trieste – GE allows a sharp distinction between phenotype, genotype and fitness, creating an unprecedented test bench (Squillero is already collaborating with prof. Medvet on these topics, see “Multi-level diversity promotion strategies for Grammar-guided Genetic Programming” (Applied Soft Computing, 2019).

A remarkable goal of this research would be to definitely link phenotype-level methodologies to genotype measures.
Expected target publications: Journals with impact factors
- ASOC - Applied Soft Computing
- ECJ - Evolutionary Computation Journal
- GPem - Genetic Programming and Evolvable Machines
- Informatics and Computer Science Intelligent Systems Applications
- IS - Information Sciences
- NC - Natural Computing
- TCIAIG - IEEE Transactions on Computational Intelligence and AI in Games
- TEC - IEEE Transactions on Evolutionary Computation
Top conferences
- ACM GECCO - Genetic and Evolutionary Computation Conference
- IEEE CEC/WCCI - World Congress on Computational Intelligence
- PPSN - Parallel Problem Solving From Nature
Current funded projects of the proposer related to the proposal:
Possibly involved industries/companies:The CAD Group has a long record of successful applications of evolutionary algorithms in several different domains. For instance, the on-going collaboration with STMicroelectronics on test and validation of programmable devices, does exploit evolutionary algorithms and would definitely benefit from the research. Squillero is also in contact with industries that actively consider exploiting evolutionary machine-learning for enhancing their biological models, for instance, KRD (Czech Republic), Teregroup (Italy), and BioVal Process (France).

Title: Accessibility of Internet of Things Devices and Systems
Proposer: Fulvio Corno, Luigi De Russis
Group website: http://elite.polito.it/
Summary of the proposal: The Internet of Things (IoT) is expected to have a significant impact on our daily lives as it will change how we interact with each other, with our objects and our spaces, be they the home or the work place. The IoT provides an opportunity to ensure equal access for everybody: e.g., interacting with the enviroment (doors and lights) through voice makes an IoT-powered space more accessible to people with physical disabilities and inclusive to many more. Furthermore, web standards for the IoT like the WebThings extend the open web platform to the physical world and provide a good starting point for making the IoT an enabler, and not a disabler. However, several challenges are still present, ranging from the interoperability of assistive technology with "new" IoT systems to accessibility support in the way we interact with and configure IoT devices, to accessibility guidelines.
This PhD proposal aims at investigating these and similar issues and to provide novel and promising strategies, frameworks, tools, and interface that can positively impact the accessibility of contemporary IoT solutions.
Rsearch objectives and methods: The objective is to assess and improve the level of accessibility of new technologies (based on IoT systems). The application areas of such systems range from smart environments (home, workplace), smart transportation, health and well-being, gaming and entertainment, as well as communication. Hundreds of new 'connected' products are being proposed by researchers, startups and industries. This proposal will focus on their interface and usability for persons with disabilities (speech impairment, cognitive impairments such as autism, motor disabilities, ...).
New smart devices, coupled with novel intelligent interaction methods, could profoundly empower persons with disabilities in their needs. However, many products are not designed with this need in mind.
One first, minor, research objective is to identify the accessibility barriers in current IoT systems and devices, both at the level of the adopted interaction methods (voice, touch, etc), and of the available user interfaces. We will identify major issues and propose alternative accessible solutions (at the algorithm or interface level). The research methods will be based on user studies, controlled experiments, interaction prototypes.
A second, major, research objective is to propose robust and applicable solutions to some of the identified challenges. We will research, design, prototype and evaluate novel IoT devices or interfaces able to solve the accessibility barriers for specific kinds of disabilities. Some examples (to be validated) are: voice assistants for persons with speech disorders, alternative physical interfaces for persons with motor disabilities, dynamic generation of user interfaces for persons with mild cognitive disabilities (including the elderly), tangible and immersive IoT-based gaming for children, etc.
The research methods will combine techniques adopted in Human Computer Interaction (such as Human centered design processes, empirical evaluations, user studies, etc) with know-how about design and realization of IoT systems (such as distributed systems, embedded devices, mobile interfaces, intelligent algorithms, etc).
Outline of work plan: The work plan will be organized according to the following four phases.
Phase 1 (months 0-6): literature review about accessibility of digital devices and interfaces; study and knowledge of IoT devices and related communication and programming practices and standards.
Phase 2 (months 6-18): assessment and analysis of accessibility of IoT devices and interfaces. In this phase, some usability assessment and testing, and some accessibility evaluation will be conducted on some significant IoT devices (e.g., conversational assistants, smart thermostats, wearable devices, …) to assess their degree of accessibility for different populations of users. The expected result is the definition of a set of accessibility issues and challenges.
Phase 3 (months 12-24): research, definition and experimentation of novel accessible interfaces for IoT systems. Starting from the results of the previous phase, some of the identified challenges will be addressed in this phase, by proposing new solutions for lowering the accessibility barriers. Such solutions could imply the design of new user interfaces (within the same device or on a different one), or improvement of the intelligent algorithms of the IoT device to compensate for user capabilities.
Phase 4 (months 24-36): the last phase will be twofold. On one hand, the work on addressing the identified challenges will be continued. On the other hand, a robust and large-scale accessible IoT system will be proposed, that includes the proposed technologies and interaction methods, and will be the basis for a comparative assessment of the attained accessibility improvement.
Expected target publications: IEEE Internet of Things Journal, ACM Transactions on Internet of Things, ACM Transactions on Accessible Computing, IEEE Pervasive Computing, IEEE Transactions on Human-Machine Systems, ACM Transactions on Computer-Human Interaction
Current funded projects of the proposer related to the proposal: One VivoMeglio project about tangible gaming.
Possibly involved industries/companies:None at the moment

Title: Automatic Generation of biofabrication pProtocols for regenerative medicine applications
Proposer: Stefano Di Carlo
Group website: https://www.sysbio.polito.it
Summary of the proposal: 3D Biofabrication is an emerging field that identifies the set of processes required for the automated generation of biologically functional products with structural organization through bioprinting or bioassembly and subsequent tissue maturation processes. Biofabrication technologies have the potential to revolutionize the Regenerative medicine and organs-on-chip (OoC) domains.
Currently, improving biofabrication protocols implies the need of executing a large number of in vitro experiments, which are costly in terms of time and resources, and come with a high risk of failure. Computational models and Design Space Exploration (DSE) techniques can support the search for new and better biofabrication protocols. Yet, this field is largely unexplored, and poses several challenges to computational approaches. Leveraging biological complexity implies, on the computational front, to deal with computational complexity and scalability issues.
This Ph.D. project aims at innovating in the domain of computational tools to support 3D biofabrication technologies, with the development of a software framework, capable of generating optimal protocols to improve the biofabrication of complex biosystems using 3D bioprinting technologies coupled with automatic culture systems.
Rsearch objectives and methods: The global objective of the project is the development of a software framework capable of generating optimal biofabrication protocols, i.e., the set of physical, biochemical, mechanical and biological stimuli that, organizing themselves in space and time, guide the starting cells towards the formation of the target biological product.

This is a combination of two main elements to be designed during the Ph.D.

The first element is a multi-scale modeling strategy allowing modeling and simulating of a full 3D biofabrication system, which includes the automated culture environment, and the target biosystem.

The knowledge necessary to model a complex biological system is often available in different formats, deriving from heterogeneous research fields. There are, at the state of the art, various models and simulators that provide important functionalities for analyzing a phenomenon but cannot be integrated together. This project aims at supporting the integration of different models and representation strategies within a holistic environment that can combine and consider them jointly.

Modeling and simulation alone are not enough to reach the goal of the project. Several protocols (i.e., sequences of stimuli) can be used to guide a biosystem towards a desired result. Design Space Exploration (DSE) algorithms simulating the reaction of the system to different sets of organized stimuli can search for optimal biofabrication protocols. This supports several stages of the biofabrication process including rapid prototyping, optimization, and integration.

The main challenge of the DSE is the size of the design space to explore. Exploring it comprehensively is computationally prohibitive due to its exponential size.
Outline of work plan: Year 1: Simulating multi-scale hybrid biological models

Computational models of biological systems are specified under a variety of formalisms. To cover biofabrication, it is necessary to include models of the intracellular mechanisms involved in biofabrication processes, as well as models of inter-cellular dynamics, biofabrication environment and stimuli. During the first year, the PhD student will design a compositional simulation approach able to combine and jointly simulate such different models on a selected use case and validate model simulations with literature data.

For this purpose, existing models need to be encapsulated into a standard "wrapper" in order to expose heterogeneous functionalities provided by third-party simulators through a set of standardized APIs.

Year 2: Exploring Design Spaces

During the second year, the PhD student will develop DSE strategies to perform over the model to search optimal biofabrication protocols. This will include researching and developing DSE algorithms devising the underlying computational complexity and developing a DSE engine to integrate into the simulation framework developed during the first year. This activity starts from the consideration that generating an optimal biofabrication protocol is similar to generating an optimized computer program. The main difference is the set of available instructions, that is constrained by the controllable parameters of the biofabrication process.

Protocol generation will be validated over data from the literature.

Year 3: Taming computational complexity

Scalability of the framework over model complexity is a killing factor in computational biology and must be seriously considered. In this domain it can be measured by looking at the simulation wall time as a function of the number of: simulated cells, iterations, available cores, equations or parameters of the model. During the third year, the PhD student will investigate this trade-off by working toward high-parallel implementations of the methods studied during the first two years able to increase the capability of handling real and complex biofabrication use cases.
Expected target publications: We expect the candidate to be able to submit her/his work to several high-level conferences (e.g., IEEE BIBM, BIOSTEC BIOINFORMATICS, etc. ) as well as top ranking journals (e.g., BMC Bioinformatics, Oxford Bioinformatics, Oxford Databases, IEEE/ACM TBIO). For the conference publication, we foreseen a minimum of two conference publications per year, while at least two publications, spawning along the second and the third year are expected
Current funded projects of the proposer related to the proposal: Poc Instrument Project: “Digital Simulation Method for the onthogenesis of a bilogical System and generation of culture protocols”.
Poc-Off “Cultura Regenerating Systems”
Possibly involved industries/companies:During the third year, the student will have a chance to collaborate with an external cellular biology lab at the University di Torino able to support the validation of the obtained results.

Title: Pervasive Information Management
Proposer: Filippo Gandino
Group website: http://www.dauin.polito.it/research/research_groups/cad_electroni...
Summary of the proposal: Pervasive technologies are composed by densely distributed, networked, low cost devices. The two main examples are wireless sensor networks and radio frequency identification. This kind of technologies involve many devices that continuously collect, elaborate and transmit data. The proper processing of this information is fundamental to exploit the actual benefits of pervasive technologies. However, this task is very complex and represents one of the main issues that are slowing their diffusion.

The research activity will be related to the whole cycle of life of the pervasive information, starting from the collection, through the transmission, up to the final analysis. The candidate will work within a team with a consolidated experience in pervasive technologies and he/she will have the opportunity to access to implemented and under development sensor networks.

The proposed research involves multidisciplinary knowledge and skills (e.g., data mining, computer network, advanced programming).
Rsearch objectives and methods: The goals of the research are related to the whole cycle of life of the pervasive information.

The first phase is the data collection. It is characterized by which data are produced, their format and how often they are collected. These characteristics are fundamental, since they affect the accuracy of the system, the power consumption and the quantity of data to store and transmit.

The second phase is the pervasive elaboration. The devices directly elaborate the collected data. They can merge more data, compute statistics, apply filters, correct values. However, the quantity and quality of these tasks must be defined by considering that the computational capability of pervasive devices are often limited like their power supply.

The third phase is the transmission. According to the previous phases, the amount of data is already determined. The frequency of the transmissions will affect the efficiency of the system update, and the power consumption.

The fourth phase is data storage. In order to effectively manage big data, the data base organization is fundamental. Moreover, a part of the data can be deleted or summarized in order to save space and improve the query efficiency.

The fifth and last phase is the data analysis and visualization. The data must be analyzed in order to find useful information. Moreover, the data analysis is important for the accuracy of the system, since the data must be calibrated, even dynamically, odd data must be removed and missing data can be inferred.

The candidate will have the opportunity to work with already working sensor networks and to cooperate in the development of new ones. Its activity will focus on the balance between efficiency and data accuracy and on the data analysis. The goals include solving specific problems and finding general strategies and techniques both in the organization of the data and in their analysis.
Outline of work plan: The candidate will start its activity by becoming familiar with the pervasive systems already developed within the research group. He will analyze the phases of the cycle of live of the information in order to improve the balance between efficiency and accuracy. During the phd, he/she will also have the opportunity to work on the design and development of new pervasive systems.
In parallel, a large part of the research activity will be focused on the last phase: the data analysis.
A relevant topic is the calibration of the data. Although there are consolidated techniques for the static calibration, they are not always applicable. The candidate will study how to calibrate data when the standard solutions are not enough. Moreover, he/she will investigate the technique for dynamic calibration, in which when two sensors meet, their calibration is updated. Starting from the state-of-the-art work, the activity will consider how to improve the accuracy also of older values according to the new calibration.
A second topic is the filtering of the data in order to delete unreliable data or to tag them as partially reliable. This task will be executed by investigating the correlation between the reliability of the data and other values (e.g., speed, humidity).
A third topic will be the inference of the missing data. This task is fundamental for data with a bi- or three-dimensional position. Moreover, with mobile sensors, not only a spatial inference but also time inference could improve the results. The candidate will investigate this topic, and will use the results also for the optimization of the node deployment.
The forth topic is the analysis of the data. Each sensor system provides different opportunities that can be reached only through effective analysis of the data. The candidate will exploit data mining and machine learning approaches.
Expected target publications: IEEE Transactions on Parallel and Distributes Systems
IEEE Transactions on Wireless Communications
IEEE Internet of Things Journal
Wireless Networks (Springer)
Current funded projects of the proposer related to the proposal:
Possibly involved industries/companies:

Title: Integration of Machine Learning and Network Virtualization for the Orchestration of Distributed Edge Infrastructures
Proposer: Guido Marchetto
Group website: http://www.netgroup.polito.it
Summary of the proposal: Recent advances in network virtualization, machine learning, and artificial intelligence have made data-driven techniques attractive to tame the increasing complexity and scale of today's network infrastructures, especially to automate network orchestration. In particular, valuable progress has been made on edge network management using learning techniques and network virtualization – i.e., Network Function Virtualization (NFV) solutions – independently. The integration of these technologies would be a further added value for network automation, but it is still challenging in several ways.
Among those challenges, data is not always available or obtainable, and even when it is, retraining machine learning models are slow and expensive. Moreover, most data analytics architectures for orchestration of edge infrastructures either ignore the wireless network's underlying dynamics or the mechanisms necessary for efficient virtualization.
The overarching aim of this research work is to propose network management solutions using machine learning with incomplete information for inference and with limited cost and training opportunities, in order to deploy scalable and efficient network automation framework within virtual edge networks.
Rsearch objectives and methods: The main objective of the proposed research is to design and implement new network management architectures that integrate machine learning and network virtualization technologies for the orchestration of edge infrastructures. The design will holistically consider inference techniques for distributed resource discovery, algorithmic solutions for machine learning driven service mapping, and distributed resource allocation decisions for (sustainable) training across wireless and wired network application processes.
In particular, the proposed solutions will integrate learning theories and network virtualization in novel ways, focusing on a few mechanisms that edge network providers run to create, deploy, and maintain valuable services: 1) inference techniques in network telemetry for resource discovery, 2) machine learning techniques for distributed network and service mapping, to optimize computation placement and transmission parameters, and 3) allocation, to bind virtual resources to the physical edge infrastructure in support of ultra-low latency networked applications.
First, the candidate will study novel solutions based on inference techniques to design network discovery protocols that can reconstruct states of a (wireless) edge network with only aggregate measurement information. Second, novel network protocols and architectures will be designed to solve network and service mapping problems with partially known network states, using unsupervised learning. Third, novel resource allocation mechanisms will be designed and implemented using distributed consensus protocols with theoretical guarantees to properly manage distributed latency-sensitive applications.
A further objective will be the evaluation of the developed solutions in significant use-case scenarios. For example, offloading of computational tasks from Internet of Things devices to the edge will be considered. In such a scenario, a proper management of the network edge resources is key to guarantee the required quality, e.g., in terms of task completion time. Other use cases might be related to telehealth or other industrial applications, where low and bounded latencies are a must for an effective service.
Outline of work plan: Phase 1 (1st year): the candidate will analyze the state-of-the-art solutions for network management, with particular emphasis to knowledge-based network automation techniques. The candidate will then define detailed guidelines for the development of architectures and protocols that are suitable for discovery, mapping and allocation within the IoT-Edge-Cloud continuum. Specific use-cases will also be defined during this phase (e.g., in the telehealth). Such use cases will help identifying ad-hoc requirements, and will include peculiarities of specific environments. With these use cases in mind, the candidate will design and implement novel solutions to deal with the partial availability of data within distributed edge infrastructures. Results of this work will likely result in conference publications.

Phase 2 (2nd year): the candidate will consolidate the approaches proposed in the previous year, considering both the network and service mapping problem and the resource allocation task. This will lead (likely as part of the third year activity) to the definition of a comprehensive and automatic network and service management framework for distributed edge infrastructures. Solutions will be implemented and tested. Results will be published, targeting at least one journal publication.

Phase 3 (3rd year): the consolidation and the experimentation of the proposed approach will be completed. Particular emphasis will be given to the identified use cases, properly tuning the developed solutions to real scenarios. Major importance will be given to the quality offered to the service, with specific emphasis to the minimization of latencies in order to enable a real-time network automation for critical environments (e.g., industrial networks or telehealth systems). Further conference and journal publications are expected.
Expected target publications: The contributions produced by the proposed research can be published in conferences and journals belonging to the areas of networking and machine learning (e.g. IEEE INFOCOM, ICML, ACM/IEEE Transactions on Networking, or IEEE Transactions on Network and Service Management) and cloud/fog computing (e.g. IEEE/ACM SEC, IEEE ICFEC, IEEE Transactions on Cloud Computing), as well as in publications related to the specific areas that could benefits from the proposed solutions (e.g., IEEE Transactions on Industrial Informatics, IEEE Journal of Biomedical and Health Informatics, IEEE Transactions on Vehicular Technology).
Current funded projects of the proposer related to the proposal: Research contract "Sviluppo di soluzioni per il monitoraggio scalabile di dispositivi di rete" with Tiesse SpA
Possible (currently under definition) unrestricted grant from Futurewei
Possibly involved industries/companies:Tiesse, Futurewei

Title: Architectures and Protocols for the Management of the IoT – Edge Computing Ecosystem for Private and Performant Data-driven Applications
Proposer: Guido Marchetto
Group website: http://www.netgroup.polito.it
Summary of the proposal: The Internet of Things (IoT) is foreseen to be a significant driver for improving the quality of everyday life, especially when paired with the edge computing paradigm. Compelling applications include industrial machine to machine communication, medical wearable devices, and smart and connected vehicles.
Managing the IoT device interaction with the edge, however, has unsolved challenges. For example, on the IoT side, often devices cannot receive software updates, yielding to indefinite vulnerabilities. On the edge computing side, instead, one challenge is to handle the large datasets generated by these devices.
The aim of this research work is to propose a set of novel network and application architectures that operates within the edge computing-assisted IoT ecosystem, contributing in three complementary ways. From the IoT side, solutions will be proposed for real-time edge-computation attestation and computation speedups. From the edge-cloud side, distributed optimization solutions will be defined to handle big data processing as data is being generated by such IoT devices. Finally, a study will be performed on the interaction between edge, cloud, and IoT with distributed learning solutions designed to regulate, optimize, and control tradeoffs between privacy and efficiency of the IoT- edge-cloud ecosystem.
Rsearch objectives and methods: In this research activity, the plan is to dissect the tussle between privacy, security, and performance of computations when handling big data workload generated by IoT devices, as well as the security, privacy, and performance of the IoT application under analysis. Using prototypes, models, and simulations, the aim is to study how to manage transmissions and computations to steer privacy and performance of the edge computing-assisted IoT ecosystem.

The theoretical contributions of the research activity will leverage optimization and distributed learning theory for the developed models, including network utility maximization problems with algorithmic guarantees. Prototypes instead will use real IoT hardware platforms, focusing on medical and geospatial applications, such as, precision agriculture.

In particular, the specific research objectives will be: (1) to propose policy-based architectures to handle the efficiency of IoT applications, that generate, offload, and consume data processed by the edge cloud. (2) Using statistical learning and network analysis, to analyze and propose solutions to detect IoT anomalies and misconfigurations to avoid data leaks and compromised devices. (3) To design and implement algorithms and network protocols that will handle big data workloads generated by these IoT devices leveraging operational research, for example, decomposition theory, as well as distributed learning techniques, such as split and federated learning, advancing the state of the art in trustworthy, private, and performant distributed edge computing, while serving medical and other IoT applications. (4) Finally, to propose a unifying architecture that will holistically consider mechanisms, policies, and constraints from the IoT side, the edge cloud side, and their interaction, to tune and optimize the networked ecosystem based on each application needs.
Outline of work plan: Phase 1 (1st year): the candidate will analyze the state-of-the-art solutions for IoT network management, with particular emphasis to the IoT-to-Cloud Continuum. The candidate will then define detailed guidelines for the development of architectures and protocols that are suitable for handling big data applications, and for the anomaly detection of the IoT and Edge ecosystem. Specific use-cases will also be defined during this phase (e.g., edge computing and AI assisted heart monitoring and medical diagnosis, as well as AI and edge computing assisted plant diagnosis). Such use cases will help identifying ad-hoc requirements of specific environments. With these use cases in mind, the candidate will design and implement novel solutions to deal with the data Collection-Computation-Consumption (CCC) workflow of many IoT application. Results of this work will likely result in conference publications.

Phase 2 (2nd year): the candidate will consolidate the approaches proposed in the previous year, advancing both solutions that are IoT driven, and solutions that augment privacy and performance of the distributed edge-cloud architecture, proposing solutions to transmission and computational resource allocation problems with algorithmic guarantees. This will lead to the definition of a comprehensive architecture for IoT application and network management, that will be fully developed in Phase 3. Solutions will be implemented and tested. Results will be published, targeting at least two journal publications, one on the IoT side, resulting from results obtained from Phase 1, and one on the edge side, resulting from work in the second year.

Phase 3 (3rd year): the consolidation and the experimentation of the proposed architecture will be completed. Particular emphasis will be given to finalize the underlying IoT-Edge network architecture, with specific customization to each identified use cases (e.g., diagnosis of humans, using available datasets). Further conference and journal publications are expected.
Expected target publications: The contributions produced by the proposed research can be published in conferences and journals belonging to the areas of networking and machine learning (e.g. IEEE INFOCOM, ICML, ACM/IEEE Transactions on Networking, or IEEE Transactions on Network and Service Management) and cloud/fog computing (e.g. IEEE/ACM SEC, IEEE ICFEC, IEEE Transactions on Cloud Computing), as well as in publications related to the specific areas that could benefits from the proposed solutions (e.g., IEEE Transactions on Industrial Informatics, IEEE Journal of Biomedical and Health Informatics, MDPI Remote Sensing or IEEE Transactions on Vehicular Technology).
Current funded projects of the proposer related to the proposal: Research contract "Sviluppo di soluzioni per il monitoraggio scalabile di dispositivi di rete" with Tiesse SpA
Possible (currently under definition) unrestricted grant from Futurewei
Possibly involved industries/companies:Tiesse, Futurewei

Title: Local energy markets in citizen-centred energy communities
Proposer: Enrico Macii, Lorenzo Bottaccioli
Group website: http://www.eda.polito.it
Summary of the proposal: A smart citizen-centric energy system is at the centre of the energy transition in Europe and worldwide. Local energy communities will enable citizens to participate collectively and actively in local energy markets. New digital tools (e.g., smart energy contracts) will be used to manage financial transactions connected to the exchange of energy among community members, different communities and with the grid. On the one side, digital and energy technology combined together will provide a framework for a more intelligent and sustainable final use of energy in buildings and cities. On the other side, citizens will need to understand how to interact with smart energy systems and local energy markets. Indeed, new complex socio-techno-economic interactions will take place in such intelligence energy systems. Given this emerging panorama, it will become even more important the understanding on the dynamics of energy technology diffusion, corporate structures of the communities, billing mechanisms and the impact that regulation and policy could have on diffusion patterns and genesis of new communities.
Rsearch objectives and methods: The diffusion of distributed (renewable) energy sources poses new challenges in the underlying energy infrastructure, e.g., distribution and transmission networks and/or within micro (private) electric grids. At the same time new roles are emerging such as prosumers, aggregators and energy communities. Actions and rules of such roles are still to defined and there is no a unique idea of implementing them. The goal of the project is to explore different corporate structures, billing and sharing mechanism inside energy communities. For instance, the use of smart energy contracts based on Distributed Ledger Technology (blockchain) for energy management in local energy communities will be studied. Moreover, the candidate will explore and identify the benefits and challenges of implementing prosumer aggregation policies. A testbed comprising of physical hardware (e.g., smart meters) connected in the loop with a simulated energy community environment (e.g., a building or a cluster of buildings) exploiting different RES and energy storage technology will be developed and tested during the three-year program. Such a testbed will be used to evaluate the potential economic benefits of different energy trading paradigms.

Hence, the research will focus on the development of agents capable of describing:
1) the final customer/prosumer beliefs desire and intention and opinions.
2) the local energy market where prosumers can trade their energy and or flexibility
3) the roles of energy aggregators as intermediaries in community energy.
4) how different tariffs and metering technologies influence differently the investment decision

All the software entities will be coupled with external simulators of grid and energy sources in a plug and play fashion.
Hence the overall framework it as to be able to work in a co-simulation environment with the possibility of performing hardware in the loop.

The outcomes of this research will be an agent-based modelling tool that can be exploited for:
- Planning the evolution of the future smart multi-energy system by taking in to account the operational phase
- Evaluating the effect of different policies and related customer satisfaction
- Evaluating the diffusion of technologies and/or energy policies under different regulatory scenarios
- Evaluating the new business model for energy communities and aggregators
Outline of work plan: 1st year. The candidate will study the state-of-the-art solution of existing agent-based modelling tools to identify the best available solution for large scale smart energy system simulation in distributed environments. Furthermore, the candidate will review the state of the art in prosumers/aggregators/market modelling to identify the challenges and identify possible innovations. Moreover, the candidate will focus on the review of possible corporate structures, billing and sharing mechanism of energy communities. Finally, it will start the design of the overall platform starting for the requirements identification and definition.

2nd year. The candidate will terminate the design phase and will start the implementation of the agents intelligence. Furthermore, it will start to integrate agents intelligent and simulators together to create the first beta version of the tool.

3rd year. The candidate will ultimate the overall platform and test it in different case study and scenarios to show all the effects of the different corporate structures, billing and sharing mechanism in energy communities.
Expected target publications: IEEE Transaction Smart Grid
IEEE Transactions on Evolutionary Computation
IEEE Transactions on Control of Network Systems
Environmental Modelling and Software
JASSS
ACM e-Energy
Current funded projects of the proposer related to the proposal:
Possibly involved industries/companies:

Title: Intermittent Computing for Batteryless Systems
Proposer: Enrico Macii, Massimo Poncino, Daniele Jahier Pagliari
Group website: http://eda.polito.it
Summary of the proposal: The claimed IoT scenario in which billions or trillions of tiny devices that compute, sense, and learn, while being also small, cheap, and operating for the entire lifetime of the things they monitor implies a batteryless future. Batteries wear out, are expensive, they are bulky, and replacing or recycling the number of batteries needed to power the billions sensors is not feasible nor environmentally responsible.

A novel paradigm called intermittent computing is now emerging, in which these nodes are powered solely by environmental sources; The devices have little or no energy storage so that energy harvesting becomes essential, supply voltages — which are normally stable and clean — become intermittent, and power failures become common events. When they occur, an appropriate mechanism to save the current execution state on a non-volatile memory must be set up.
This mode of operation has implications in every phase of their design: architectures, hardware components, software implementation and communication.

This research project focuses on the exploration of all the aspects above in order to deploy sophisticated intermittent computing solutions.
Rsearch objectives and methods: The research objective is to develop a set of techniques that address the various issues involved in intermittent systems, such as:

- Exploitation of the entire memory hierarchy of the system during backups; some data have freshness requirements that allow them to be stored on SRAM or even DRAM, others need to be backed up on the NVM.

- Development of software/hardware modules for predicting environmental quantities. Having a rough estimation of the available energy can drive decisions on future backups. Since the duration of on-off intervals of these systems is small (in the order of ms), environmental predictions, which have larger time constants (seconds), will serve as long-term estimates.

- Hardware and energy-friendly estimation the off time of the system. This is an essential facility, as during off times the system is not powered so no timer can be used.

- Intermittent-aware software. This will either encompass (1) code instrumentation to support state recovery after a power-off, or (2) re-design of algorithms to provide approximate or partial outputs when interrupted.

- Energy and functional simulators for the whole system. This is essential for a quick prototyping and what-if analysis of the various system parameters.

While there are some research works addressing some of these issues, they suffer from two limitations. First, most of them are strongly focused on execution correctness and put less emphasis on energy. Secondly, they mostly rely on abstract models of the architecture, of the computation and of the environmental sources, and lack validation.

For this reason, one strength of our research will be the deployment and testing of the proposed techniques on real hardware platforms such as the TI CC2650. This will allow us to provide an accurate assessment of the impact of the proposed techniques in terms of execution correctness and device lifetime.
Outline of work plan: 1st year. The candidate will study state-of the-art techniques in particular i) existing architectures for intermittent systems, ii) existing strategies for the energy management of these systems both in hardware and software, iii) existing simulation platforms, and iv) existing commercial platforms. Before the end of the 1st year a target system will be identified that is both available as a real platform and for which a simulator is available. Moreover, the target open research problems will be identified. Planned publications: 1 conference paper.

2nd year. Based on the outcomes of the first year, specific techniques will be evaluated on the target platform, first by simulating them and porting the most promising ones on the physical platform. Planned publications: at least 1 or 2 conference papers.

3rd year. The methodology and the algorithms developed in the previous years will be validated to prove their robustness and scalability in being applied on realistic workload to be executed on the platform. Planned publications: 1 journal paper.
Expected target publications: IEEE Transactions on CAD
IEEE Transactions on Computers
IEEE Journal on Internet of Things
IEEE Transactions on Circuits and Systems (I and II)
IEEE Design and Test of Computers
IEEE Sensors Journal
ACM Transactions on Embedded Computing Systems
ACM Transactions of Design Automation of Electronic Systems
Current funded projects of the proposer related to the proposal: - End-to-end digitalised production testbeds (EIT X-KIC)
- AMBEATion (H2020, MSCA-RISE)
- MADEIN4 (H2020, ECSEL)
- DTWIN (Regional Funds)
Possibly involved industries/companies:- STMicroelectronics
- Reply

Title: Automatic hardware-aware design and optimization of deep learning models
Proposer: Enrico Macii, Massimo Poncino, Daniele Jahier Pagliari
Group website: http://eda.polito.it
Summary of the proposal: Deep learning models are embedded in many emerging applications, from computer vision to industrial assets monitoring. However, their excellent accuracy comes at the cost of high computational complexity. Therefore, optimizing model architectures, i.e., selecting appropriate hyperparameters, is of paramount importance to achieve the required accuracy at the minimum latency, memory, and energy cost. At different scales, this is critical both for high-performance cloud platforms, and for energy-constrained edge devices. Additional complication is given by the fact that both edge and cloud devices are increasingly heterogeneous, mixing CPUs with GPUs, accelerators and FPGAs. Thus, a model optimized for one platform may not be optimal for another target.

Classically, hyperparameters were tuned manually, often resulting in suboptimal choices. Recently, Neural Architecture Search (NAS) algorithms have emerged as a systematic alternative. Despite its increasing popularity, NAS is still an open research field. Two critical issues are related with reducing the search complexity (to avoid huge runtimes) and taking into account hardware heterogeneity, by embedding a hardware model in the search.

This project focuses on exploring novel NAS approaches, with particular focus on: (i) extending efficient NAS solutions to new types of models (e.g., TCNs, Transformers, etc.) and (ii) improving their hardware-awareness.
Rsearch objectives and methods: The research objective is to extend and improve existing tools for the automatic optimization of deep learning models on heterogeneous hardware. In particular, we envision the study of a number of techniques focusing on two key aspects:

- Complexity reduction: early NASes relied on complex meta-heuristics or reinforcement learning, requiring the evaluation of a large number of solutions, and resulting in huge runtimes. More recently, differentiable NAS (DNAS) have been proposed to reduce such complexity. Nonetheless, DNASes are less flexible and can explore smaller solution spaces. A first objective will be to extend efficient DNASes to larger design spaces, adapting them to explore new hyperparameters and new types of models (e.g. Temporal Convolutional Networks, Transformers, etc.) besides the standard Convolutional Neural Networks for which they are typically designed.

- Hardware awareness: Another limitation of DNASes is being constrained to using simple (and often inaccurate) models for the computational cost of a network, which cannot reflect the actual latency or energy cost of a given hyperparameters combination, especially due to hardware heterogeneity. A second objective will be to develop accurate hardware-aware cost models to be integrated in DNASes. For high-performance parallel platforms, these should account for the characteristics of the cores and accelerators available on chip, as well as for the data transfer costs between cores and through the memory hierarchy. For edge devices, cost models should be energy-oriented and consider effects associated to the power envelope and to the battery non-idealities.

While there are research works addressing these issues, they mostly focus on CPUs and GPUs only. Fewer works concentrate on accelerator-rich heterogeneous systems, or on “tiny” edge devices. One focal point of our research will be the validation of all proposed techniques on real hardware platforms, such as Xilinx’s Zynq, GreenWaves’ GAP8, NVIDIA’s Jetson TX2, etc.
Outline of work plan: 1st year. The candidate will study state-of-the-art on deep learning, on NAS and on heterogeneous hardware. Precisely, the candidate will survey: i) the rapidly changing state-of-the-art of deep learning architectures ii) the existing NAS algorithms, based on meta-heuristics, reinforcement learning and DNAS. iii) the state-of-the-art on heterogeneous SoCs, both for high-performance and energy-efficiency. Before the end of the 1st year, the candidate will identify the main limitations of current DNASes, and will become familiar with a set of hardware platforms on which to evaluate the proposed improvements. Planned publications: at least 1 conference paper.

2nd year. Based on the outcomes of the first year, specific NAS improvements will be designed and new hardware-aware cost models will be developed, focusing on the selected target platforms as benchmarks. The effectiveness of the proposed techniques will be assessed both through system simulations and through the deployment of individual networks on real hardware. Planned publications: at least 1 or 2 conference papers.

3rd year. The different techniques and cost models developed in the previous years will be integrated, ideally into a single configurable NAS tool able to support multiple types of network and heterogeneous hardware. Moreover, the end-to-end results of the proposed solutions will be assessed by deploying entire applications on the target platforms. Planned publications: at least 1 journal paper.
Expected target publications: IEEE Transactions on CAD
IEEE Transactions on Computers
IEEE Journal on Internet of Things
IEEE Transactions on Emerging Topics in Computing
IEEE Transactions on Neural Networks and Learning Systems
IEEE Design and Test of Computers
ACM Transactions on Embedded Computing Systems
ACM Transactions of Design Automation of Electronic Systems
Current funded projects of the proposer related to the proposal: - End-to-end digitalised production testbeds (EIT X-KIC)
- AMBEATion (H2020, MSCA-RISE)
- MADEIN4 (H2020, ECSEL)
- DTWIN (Regional Funds)
Possibly involved industries/companies:

Title: Simulation and Digital Twin development for Industrial Cyber Physical Systems
Proposer: Enrico Macii, Sara Vinco
Group website: https://eda.polito.it/
Summary of the proposal: The Industry 4.0 paradigm requires to tightly monitor physical systems, ranging from production plants to wind turbines, with the goal of enhancing monitoring, control and optimization of their operation. This requires the development of models of the physical systems, that will be tightly connected with the cyber infrastructure to allow run time monitoring and reaction. Such a coupled interdependency between cyber (that gathers data from the environment to apply algorithms) and physical (that reacts to the decisions of the cyber portion) goes under the name of digital twin. The goal of this project will be developing models of physical behaviors, ranging from energy harvesters to industry equipment, to analyze the modeling approaches and allow the application of energy-efficient algorithms and solutions. The models will be of different types, mostly constructed with two complementary approaches: top-down, with formal and physical descriptions of equipment behavior, or bottom-up, with the application of machine learning algorithms that extrapolates information from sensed data to predict future evolution of the physical system. The joint application of such techniques, together with the other Industry 4.0 enabling technologies (big data, IoT, …) allow to foresee the state of health of the physical system, apply energy optimizations and estimate its future evolution.
Rsearch objectives and methods: The objectives of the Ph.D. plan are the following:
- Developing the competences on the physical systems modeling, including machine learning approaches and simulation languages like SystemC-AMS and Simulink
- Identifying modeling solutions available at state of the art suitable for the identified application scenario to make a critical analysis of the current solutions and of their limitations
- Providing top-down and bottom-up modeling solutions for a specific application case study, to allow energy efficiency optimization and state of health improvement of the monitored system, identifying a scalable and effective trade off between simulation accuracy and real time simulation speed
- Application of CAD languages and methodologies to a new application scenario, with extensions of the available solutions
The identified solutions can be used to:
- Optimize the energy consumption of the monitored system, by optimizing the production settings and the production recipes that control physical plant evolution
- Monitor and enhance the state of health of the physical system, through the early identification of unexpected behaviors
Outline of work plan: 1st year. The candidate will study state-of the-art techniques to simulate physical systems, including i) the main simulation languages (SystemC-AMS, Simulink) and ii) machine learning frameworks adopted in the context of physical plant monitoring. The candidate will identify a case study of interest and study the current strategies available in the literature about available models and optimization results. The case study will be selected among a number of scenarios defined by projects involving industrial partners.

2nd year. Based on the outcomes of the first year, the candidate will identify the suitable models for the identified case study, and apply both the top-down and bottom-up modeling strategies to the domain of interest, to achieve effective monitoring. This will require an Introduction to the enabling technologies of Industry 4.0 enabling technologies (big data, IoT) to enable a successful interaction with the monitored physical system. This will allow the publication of conference papers to modeling conferences and the submission of one journal paper.

3rd year. The methodology and the algorithms developed in the previous years will be validated to optimize the evolution of the physical system, both in terms of energy consumption and of state of health preservation, to achieve lifelong improvement of the system. The focus will thus be on the application of overall strategies for monitoring the system state of health and optimization of production and of energy efficiency. This will lead to the publication of at least one journal paper.
Expected target publications: The work developed within this project can be submitted to conferences (DATE, DAC, …) and to journals like:
- IEEE Transactions on Computers
- IEEE Transactions on CAD
- IEEE Transactions on Industrial Informatics
- ACM Transactions on Embedded Computing Systems
- ACM Transactions of Design Automation of Electronic Systems
Current funded projects of the proposer related to the proposal: H2020 Serena
H2020 Manuela
H2020 Mesomorph
Possibly involved industries/companies:FCA, Comau, Prima Industrie

Title: Algorithms, architectures and technologies for ubiquitous applications
Proposer: Renato Ferrero
Group website: http://www.cad.polito.it/
Summary of the proposal: Ubiquitous computing is an innovative paradigm of human-computer interaction, which aims at the integration of technology into everyday objects to share and process information. It envisions accessibility of computing and communication services every time and everywhere, as well as calm technology, which asks minimal attention to the user when interacting with the system. Current technology trend is moving from one side towards smart devices, which offer mobile, complex and personalized functionalities, and from the other side towards embedded systems disappearing in the physical world and performing simple and specific tasks. It follows that several issues must be addressed in the development of ubiquitous applications: architecture design for interlinking the ubiquitous components (smart devices and embedded systems); integration of different technologies (e.g., sensor networks for monitoring physical conditions, RFID network for tagging and annotating information, actuators for controlling the environment); development of algorithms for smart and autonomous functionalities. The research activity of the PhD candidate regards the design, development and evaluation of ubiquitous applications, so he/she is required to own multidisciplinary skills (e.g., distributed computing, computer network, advanced programming).
Rsearch objectives and methods: The research objectives concern the identification of requirements and the investigation of solutions for designing and developing ubiquitous applications. More in details, the following objectives will be pursued:
1) to develop distributed architectures able to support both local and remote services, thus increasing the number of functions offered and avoiding duplication of code. Resource availability and access (storage, applications, data for end-user) will be based on cloud computing and fog computing.
2) to develop context-aware applications, which can identify the most useful services for the user, instead of proposing the full list of available functionalities. This can limit the user's effort in interfacing the system and can reduce the computational requirements. The involved technology consists of wireless sensor networks, which provide useful information about the physical environment (e.g., location, time, temperature, light…), and RFID networks, for context-based query, item location and tracking, automated access to physical place or virtual service.
3) to enhance the autonomy of the ubiquitous system. For example, algorithms for power savings will increase the lifetime of mobile components, thus reducing the need of human maintenance.
4) to develop "smart" systems for proactive behavior and adaptability in dynamic contexts. The research will focus on modeling the physical environment, as well as human behavior. Limitations are due to the dynamicity of the environment, its incomplete observability, the impossibility to completely determine the user actions and goals. Therefore, algorithms for handling the incompleteness of the system and the non-deterministic user behavior will be designed.
Outline of work plan: The PhD research activities are organized in two consecutives phases.
In the first phase the PhD candidate will improve his/her background by attending PhD courses and by surveying relevant literature, then he/she will apply the learnt concepts to tackle specific issues in the implementation of ubiquitous systems. In particular, the first research hints will regard applications already developed by the research group, such as air pollution monitoring and thermal monitoring in smart building: the PhD candidate may be in charge of enhancing the ubiquity of applications developed in this context. As already existing applications will be considered in the first phase, the research objectives can be considered as independent from each other. In this way, it is possible to anticipate the expected outcome (personal scientific contribution, research papers) by focusing on one research objective at a time, since there is no need to master all concepts in advance.
In the second phase, when the training is completed and the PhD candidate owns a full vision of the matter, he/she will be able to evaluate the ubiquity of existing solutions, in particular with respect to the automation of human tasks and the access to information anywhere and at any time. He/she will be able to propose technologic improvements, and/or design new solutions for solving problems within the research group expertise. The PhD candidate will be involved in research projects aiming at designing and implementing new ubiquitous applications. He/she will be required to exploit his/her competence to analyze and solve real problems and finally to evaluate the performance of proposed solutions.
Expected target publications: - IEEE Pervasive Computing
- IEEE Journal on Internet of Things
- ACM International joint Conference on Pervasive and Ubiquitous Computing (UbiComp)
- IEEE International Conference on Pervasive Computing and Communications (PerCom)
Current funded projects of the proposer related to the proposal: Two proposals (PRIN 2020 and Horizon 2020 Green Deal) concerning similar research topics have been submitted in January, 26th.
Possibly involved industries/companies:

Title: Gamification of E2E Software Testing
Proposer: Marco Torchiano, Luca Ardito
Group website: https://softeng.polito.it
Summary of the proposal: System testing through the Graphical User Interface (GUI) is a valuable form of Verification and Validation for modern applications, especially in graphically-intensive domains like web and mobile applications. However, the practice is often overlooked by developers mostly because of the associated costs and missing live feedback about the quality of defined test sequences.

The PhD Research Proposal aims at building on top of existing Gamification approaches for Software Engineering, to define gamified approaches to GUI testing, metrics for test sequence evaluation, and tools to apply them in research and industrial settings.

The candidate will perform a systematic analysis of the state-of-the-art of gamified approaches to Software development and testing, define open-source software modules to enable their application to component-based and visual GUI testing, and perform empirical evaluations of the feasibility and of the advantages of the application of the paradigm in real-world settings.
Rsearch objectives and methods: Evidence from Software Engineering literature suggests that the testing phase is often underestimated by programmers as it is considered a boring, costly and repetitive task, as well as neglected in terms of time and resources. Despite that, testing is a fundamental phase of software development, to avoid error-prone code and application misbehavior, and to reveal defects before release to the final users.

A crucial testing facet for modern domains is System testing conducted through the Graphical User Interface (i.e., GUI testing), focusing on the visual interaction with the tested apps. In GUI testing, the tester has to interact with the system only via its Graphical User Interface. Testing with this approach is important especially when the product to develop is a website or a smartphone app, in which the visual aspect is in continue evolution and most of the interaction with the users is conducted through the GUI itself.

Gamification, which consists in the use of elements, philosophies and mechanics that are typical of game design in non-playful contexts, can help testers to perceive the testing process in a more entertaining and engaging way, whilst enhancing their performance in terms of efficacy and efficiency of the generated test sequences.

The objective of the PhD Research Proposal is thereby to define novel approaches that enable gamification of test definition for mobile and web GUI testing, with the goal of improving efficacy and effectiveness of test suites.
Outline of work plan: While model-based testing is able to cover most interactions at least in principle, it is difficult to cover meaningful or otherwise real-world E2E scenarios. A radically different approach, such as exploratory testing, brings the human creativity in the loop but it does not constitute a systematic approach. A gaming approach can represent a bridge between the two above methods. The gaming approach poses two fundamental challenges: both maximizing the efficacy of tests (e.g. in terms of coverage) and keeping the players engaged.

From a methodological perspective the PhD Research Proposal covers the following steps:

(1) the definition of a set of flexible scoring schemes to guarantee high quality of test suits,
(2) the adoption of models -- both of the UI and of the underpinning business processes -- in order to support the scoring schemes,
(3) the definition of crowdsourcing-based methodologies to enable contribution of multiple testers to the generation of test sequences at the same time;
(4) the assessment of the benefits provided by the application of gamification concepts, through:
a) Empirical evaluations of the efficiency of generated test sequences;
b) Experimental assessments of the perception of the practice by testers;
c) Industrial case study to evaluate the feasibility of the technique in real-world scenarios.

From a technological platform point of view a toolset will be developed, to adequately support the gamified test definition activities.
Expected target publications: The target for the PhD research includes a set of conferences in the general area of software engineering (ICSE, ESEM, EASE, ASE, ICSME) as well as in the specific area of testing (ICST, ISSTA).
More mature results will be published in software engineering journals, the main being: IEEE Transactions on Software Engineering, ACM TOSEM, Empirical Software Engineering, Journal of Systems and Software, Information and Software Technologies.
Current funded projects of the proposer related to the proposal:
Possibly involved industries/companies:Industries/companies that are involved in the proposal Several companies expressed concerns on mobile UI tests, among them three showed more interest in the specific goals of this PhD proposal: Reale Mutua Assicurazioni, Intesa SanPaolo, Engineering

Title: Stimulating the Senses in Virtual Environments
Proposer: Fabrizio Lamberti
Group website: http://grains.polito.it/
Summary of the proposal: The goal of virtual reality (VR) and augmented-reality (AR) is to immerse users in computer-generated, virtual environments (VEs) which could be either fully or partially digital. Interacting with VEs poses challenges which are very unique to these technologies. In fact, an ideal VE shall let the users feel as they are physically performing a task. However, sense of immersion and presence in current VEs are still far from ideal.

Notwithstanding, there are evidences of the fact that, in order for, e.g., a VR/AR-based simulation or a training to be effective, quality of graphics, of sound, of haptic feedback, etc., do not necessarily have to be realistic, just credible. It might not be essential to recreate every possible stimulus provided from the real world, every component of the task, every sensation.

Depending on the task, level of “fidelity” required could vary significantly. A number of experiments have been carried out in various domains to evaluate actual requirements, but their heterogeneity makes the applicability of available findings to further scenarios extremely hard to achieve.

This proposal is to dig into this research field, and define a methodology for identifying (and supporting) the level of realism required for the experience at hand.
Rsearch objectives and methods: Over the past years, a number of techniques have been proposed to boost the level of realism of digital contents populating both VR and AR-powered immersive VEs. The aim has been to improve the way users perceive the VE by simulating an increasing number of senses, by both enhancing fidelity in reproduction of sight and hearing, as well as developing new approaches to simulate touch, smell and taste.

In order to achieve a complete sense of immersion and presence, every stimulus provided from the real world, every sensation shall be possibly recreated, by making interaction as much faithful as possible. However, technology is still not able to stimulate all the users’ senses in a suitable way.

To deal with these limitations, researchers started to study to what extent the quality of contents and of the interactions users can have with them can influence the effectiveness of the task to be performed. Although, in some cases, findings obtained by experimenting with a given technology and/or with a particular task may generalize, in most of the cases results are specific to a given application domain (health, industry, etc.) or task (such as, e.g., surgery preparation, rehabilitation, training, for health).

The goal of this research will be to explore the large design space of simulation fidelity from many possible perspectives.

The limit of various technologies will be investigated, and approaches to cope with them proposed, by considering both professional as well as consumer settings. Furthermore, a wide set of tasks encompassing a representative set of application scenarios will be considered, with the aim to identify best practices and guidelines applicable across domains.

To this aim, applications with a training purpose will be particularly investigated through user studies since, besides qualitative indicators, quantitative metrics could be obtained as well by focusing on learners’ performance.
Outline of work plan: During the first year, the PhD student will review the state of the art in terms of techniques/approaches developed/proposed to deal with the issue of (level of) fidelity in VR- and AR-based environments. Afterwards, he/she will start to investigate new methods for dealing with content- and interaction-related quality issues, experimenting with technologies that are continuously appearing on the market. He or she will then apply devised solutions to specific use cases provided by funded research projects managed by the GRAphics and Intelligent Systems (GRAINS) group and the VR@POLITO lab, in collaboration with national and international companies and institutions. Domains of interest could encompass, e.g., energy, healthcare, robotics, autonomous systems, emergency management, etc. Results of these activities will be summarized in one or more publications that will be submitted to conferences in the field. The student will complete his/her background in AR, VR and human-machine interaction (HMI) by attending relevant courses.

During the second and third year, the work of the PhD student will build onto a few simulation scenarios related to one or more application domains (possibly selected among those mentioned above and/or related to challenges proposed by involved companies/institutions) with the aim to identify specific research questions/stress limitations concerning fidelity and, hence, design suitable solutions advancing the state of the art in the field. Activities will encompass the evaluation of users’ reaction to varying level of fidelity under different conditions, with the aim to build a set of measurable indicators and draft a collection of principles/guidelines to be considered in the design of next-generation immersive VEs. Results obtained will be reported into other conference works plus, at least, a high-quality journal publication.
Expected target publications: The target publications will cover the fields of computer graphics and human-machine interaction. International journals could include, e.g.:

IEEE Transactions on Visualization and Computer Graphics
IEEE Transactions on Human-Machine Systems
IEEE Transactions on Emerging Topics in Computing
IEEE Transactions on Learning Technologies
IEEE Computer Graphics and Applications
ACM Transactions on Graphics
ACM Transactions on Computer-Human Interaction

Relevant international conferences could include ACM CHI, IEEE VR, ACM SIGGRAPH Eurographics, etc.
Current funded projects of the proposer related to the proposal: Topics addressed in the proposal are strongly related to those tackled in the following projects managed by the proposer (and/or other possible projects under evaluation):

- E2DRIVER (EU H2020), on the use of VR for training in industrial settings.
- PITEM RISK FOR and PITEM RISK ACT (EU ALCOTRA), on the use of VR for training in emergency situations in trans-national scenarios.
- Research grant from LINKS on the creation of VR tools for training operators on Chemical, Biological, Radiological and Nuclear (CBRN) defense.

Activities will be carried out in the context of the “VR@POLITO” initiative and its laboratory (hub) at the Department of Control and Computer Engineering (DAUIN).
Possibly involved industries/companies:- KUKA Robotics
- LINKS
- Aeronautica Militare, 3° stormo Villafranca di Verona
- Regione Piemonte (Protezione Civile, AIB)

Title: Context and Emotion Aware Embodied Conversational Agents
Proposer: Andrea Bottino, Fabrizio Lamberti
Group website: https://areeweb.polito.it/ricerca/cgvg/projects.html
Summary of the proposal: Recent advances in Machine Learning and Artificial Intelligence have resulted in a growing interest in using Embodied Conversational Agents (ECAs) in Human-Computer Interaction (HCI). ECAs are animated virtual characters capable of simulating a human-like face-to-face conversation using natural language processing (NLP) and multimedia communicative behaviours that include verbal and non-verbal clues. The availability of increasingly powerful and connected sensors allows ECAs to access contextual information and interact autonomously with human beings and the environment. The possibility of leveraging a virtual body and voice for interaction requires researchers to include social-emotional components in ECA behaviour design. In other words, these agents should have a personality, emotions, and intentions and the capability to express them with voice, hands, head, and body movements. Simultaneously, they should enhance their social interaction with the user by simulating and triggering empathy. Given their capabilities, ECAs have the potential to play an increasingly important role in a plethora of applications ranging from educational and training environments, health and medical care, virtual assistants in industry, and virtual companions in games. Despite that, improving the effectiveness of ECAs requires substantial contributions from the research. In particular, ECAs should be made simple to design and implement, capable of fully expressing and conveying (believable) emotions, and leveraging the (fine-grained) analysis of human affects to create a robust empathic bond with the end-users.
Rsearch objectives and methods: Implementing ECA is a cumbersome and complicated process, which requires taking into account several different elements (NLP, context sensing, emotion modelling, affective computing, 3D animations), which, in turn, involve specific technological and technical skills. Thus, there is the need to develop a simple framework that (i) allows developers to support the rapid design, prototyping and deployment of ECA in a variety of heterogeneous use cases and (ii) is easily extensible, enabling the introduction of novel features that expand the current capabilities of ECAs. This framework will be the first relevant outcome of the work.

Then, the research project aims at tackling several issues that can improve the quality and effectiveness of the end-user experience when they interact with ECAs. In particular, it aims at addressing the following problems.

A substantial portion of our communication is nonverbal. Every day, we respond to thousands of nonverbal cues and behaviours, including postures, facial expressions, eye gaze, gestures, and tone of voice. However, some of these clues are currently relatively little-used in this context. Affective AI provides robust methods to detect facial expressions from off-the-shelf cameras and in the wild. However, few works have been done to exploit user postures or specific gestures (e.g., nodding can express agreement, and a frown can convey disapproval or disagreement).

The extraction of paralinguistic factors such as tone of voice, loudness, inflexion, and pitch can provide information about the actual emotional states of the other peer in the communication. Thus, computational mechanisms capable of extracting these variables from the user's voice analysis are sorely needed. The same paralinguistic factors should be available to modulate the ECA response according to its emotional states. On the contrary, one of the problems with present text-to-speech libraries is that they pronounce everything with the same tone, making it impossible to communicate feelings through voice. One of the possibilities that the research will explore is the use of style-transfer approaches, for instance, similar to the ones implemented in the visual contexts for transferring the style of a painting on a photo taken with a digital camera. We expect these approaches to "transfer" recognizable emotional forms from a set of reference samples to the synthesized utterance.

The general approach in managing voice interaction between the user and the ECA is to define a set of intents, each associated with a training set of utterances for recognition. However, identifying intents and creating a suitable training set is complicated and tedious. Thus, a relevant contribution to soften this issue would be developing automatic or semi-automatic approaches for harvesting, mining, and clustering textual corpus related to the specific scenario addressed.

Finally, the project foresees the development of advanced Computer Vision and Machine Learning approaches to allow ECAs to exploit visual information for capturing users' affect and actions and use these pieces of information to proactively respond to users' intents and improve the ECA empathic and emotional response.
Outline of work plan: In the first year, the project will start from a systematic review of the current state of the art of ECAs. The outcome of this works will serve to analyze the current limitations, identify promising approaches, and inform the design and development of the solutions to address the problems addressed by the research proposal. The candidate will also start designing, developing, and assess a framework that allows the rapid prototyping and deployment of ECA in different contexts.
In the following years, the candidate will develop solutions addressing the research mentioned above objectives. These solutions will aim at extending the capabilities offered by this framework and, ultimately, the features that ECAs can exploit to improve the quality and effectiveness of the interaction with the end-users. The proposed techniques will be applied to relevant use-cases using VR and AR applications and validated through user studies involving panels of volunteers.
Expected target publications: International peer-reviewed journals in the fields related to the current proposal, such as IEEE Transactions on Affective Computing, IEEE Transactions on Visualization and Computer Graphic, IEEE Transactions on Human-Machine Systems, ACM Transactions on Graphics, IEEE Transactions Image Processing, IEEE transactions Pattern Analysis and Machine Intelligence, Pattern Recognition, Pattern Recognition Letter, Computer Vision and Image Understanding. Relevant and well-reputed international conferences, such as IEEE Face and Gesture Recognition (FG), ACM Intelligent Virtual Agents, IEEE VR, CVPR, Eurographics.
Current funded projects of the proposer related to the proposal:
Possibly involved industries/companies:

Title: Robotic 3D Vision for Scene Understanding Across Domains
Proposer: Tatiana Tommasi
Group website: https://www.dauin.polito.it/personale/scheda/(nominativo)/tatiana...
Summary of the proposal: 3D cameras are crucial for many robotic tasks as object detection, pose estimation and overall scene understanding. Depth images have lower complexity than their twin rgb images because they miss color and texture and are generally considered more robust to visual domain variations (eg illumination conditions). However, since the advent of Microsoft Kinect, many other depth sensors have been developed and artificial agents have been endowed with different systems depending on the specific requirements. Due to the variety of acquisition logics (structured light, time-of-flight, active stereo), the obtained depth data differ significantly which still limits the possibility to export knowledge and learned models across different platforms. The aim of this thesis is to run an extensive analysis on the current existing multi-cue (rgb + depth) deep learning models to check how their performance is affected when changing the depth data domain. We will design tailored domain adaptation strategies to alleviate the existing distribution shift and allow knowledge transfer.
Rsearch objectives and methods: Robotics systems need to understand their environment to navigate and interact with the surroundings. RGB-D cameras use range imaging technologies to pair the depth geometric information with color and texture of rgb images. Currently RGB-D samples are the most effective input data to develop 3D space learning models for robotic vision, however both the considered visual modalities may suffer for domain shift. The problem of closing the domain gap among different rgb domains has been extensively studied [1], but less attention has been dedicated to the depth modality. A recent work has highlighted the presence of a significant domain gap between synthetic and real depth images [2], but the proposed solution is mainly focused on object classification and does not consider the shift due to different real cameras. Other works deal with semantic segmentation and synthetic to real domain shift, but the depth is only used as a privileged information for the synthetic source domain and is not available for the real world target [3].
This thesis aims at broadening the study on depth and multi-cue domain adaptation for robotics scene understanding. We will go beyond object classification, considering the tasks of scene recognition [4], object detection [8] and pose estimation [5] from images and videos [6]: how they are affected by changing the depth information collecting device, and how to avoid the cross-domain performance drop by introducing adaptive models.

[1] A Survey of Unsupervised Deep Domain Adaptation, ACM T-IST2020
[2] Unsupervised Domain Adaptation through Inter-modal Rotation for RGB-D Object Recognition, RAL2020
[3] DADA: Depth-Aware Domain Adaptation in Semantic Segmentation, ICCV2019
[4] Translate-to-Recognize Networks for RGB-D Scene Recognition, CVPR2019
[5] YCB-M: a multi-Camera RGB-D dataset for object recognition and 6dof pose estimation, ICRA2020
[6] https://nikosuenderhauf.github.io/roboticvisionchallenges/scene-understanding
[7] Single Image Depth Perception in the Wild, NIPS2016
[8] SF-UDA-3D: Source-Free Unsupervised Domain Adaptation for LiDAR-Based 3D Object Detection, 3DV2020
Outline of work plan: M1-M6: Definition of the experimental testbed. Most of the existing scene recognition datasets were obtained collecting images regardless of the used rgbd camera (eg SUN-RGBD [4]). By re-organizing the dataset we can obtain new domain adaptation and generalization settings. Literature review and implementation of existing baselines. Metric choices and extensive benchmark evaluations.

M6-M12: Implementation of an end-to-end fusion model rgb-to-depth and depth-to-rgb that extends [4]. The self-supervised modality translation model can be used to close the domain gap providing more details than image rotation in [2]. Testing on defined benchmarks. Writing scientific report on findings of Y1.

M13-M24: Besides useful for scene recognition, the model proposed in Y1 is inherently generative: it can work for depth prediction in the wild [7]. Besides evaluating this scenario, we will also consider pose estimation and object detection tasks studying how to refine the model in those cases [5]. Assessment of work on the established benchmarks. Writing scientific report on findings of Y2.

M25- M36. Moving the deep architecture obtained in Y2 from still images to videos and slam tasks for navigation and mapping [6]. Assessment of work on the established benchmarks. Writing scientific report on overall findings of the project. Writing of PhD thesis.
Expected target publications: It is expected that the scientific results of this thesis will be reported in the top conference in the field of computer vision (IEEE CVPR, IEEE ICCV, ECCV) and robotics (ICRA, IROS, RSS). At least one journal publication is expected on one of the following international journals: IEEE PAMI, IJCV, CVIU, RAL.
Current funded projects of the proposer related to the proposal:
Possibly involved industries/companies:Ferrero SpA

Title: Proposal title (max 150 characters) Automated Painting: Object Modeling, Trajectory Learning and Adaptation
Proposer: Tatiana Tommasi
Group website: http://www.tatianatommasi.com
Summary of the proposal: Spray painting is a task commonly performed by robots in industry. Automating this process offers advantages on the consistency of the results and repeatability over even the most skilled human operator, with the further benefit of limiting human exposure to hazardous environments. The goal of this thesis is to reduce human intervention in the learning process of the robotic arm motion path. The tip of the spray gun should follow the trajectory that produces the highest painting quality with uniform material distribution and minimum wasted paint. We will focus on solving this optimization process decomposing it in three parts: (1) modeling the object to be painted (2) learn the trajectory from pre-defined ground truth paths or imitation via a painting quality based reward (3) adapting the model to new scenarios (from CAD simulators to reality), new objects and new painting modalities (eg protection/esthetic painting, coating).
Rsearch objectives and methods: (1) To paint an object it is first of all necessary to get its 3d representation. Often complex strategies are used for this acquisition process (eg barrier sensors [1]), while our study will focus on the use of one or multiple RGB-D cameras. We will analyze how to position the cameras both with fixed allocations and on the moving robotic arm. Once the object is captured and its exact position detected, a machine learning model will be (learned and then) applied for its segmentation. Indeed, decomposing the object in convex parts allows to simplify the painting task and its transfer to new objects.

(2) The painting trajectory can be learned starting from the object parts, their ground truth trajectories and related painting quality score. The task can be formulated as a generative problem (predict a line given a surface, while fixing all the space/time dynamic information of the arm-fixed distance from the object, uniform velocity, orientation orthogonal to the object surface). Alternatively it can be a reinforcement / imitation learning task [2] with the state described by all the tip information (3d space position, orientation, velocity) and maximizing the reward based on painting quality score. We will investigate the effectiveness of both these formulations.

(3) Trajectory generation is mainly studied on simulated data, however the final model needs a further adaptation step to be applied on the real world images. We will finally study how to close the synthetic to-real domain gap and also analyze the extension of the considered model in case of new objects and new painting scenarios. For this purpose we will exploit style transfer, meta-learning and self-supervised learning inspired by [3,4,5].

[1] Automatic Path and Trajectory Planning for Robotic Spray Painting, Robotik2012
[2] One-Shot Imitation Learning, NeurIPS2017
[3] RL-CycleGAN: Reinforcement Learning Aware Simulation-To-Real, CVPR2020
[4] One-Shot Imitation from Observing Humans via Domain-Adaptive Meta-Learning, RSS2018
[5] Self-Supervised Policy Adaptation during Deployment, arXiv2020
Outline of work plan: M1-M6: This initial period will be dedicated to literature review and to the definition of the experimental setup. The student will take confidence both with the simulated and real environments for robotic spray painting. This initial study will also finalize the RGB-D camera setting to then perform an extensive real world data collection for objects and painting trajectories.

M6-M12: Hands-on experience on existing 3D object detection and segmentation methods to choose the reference baselines. A dedicated segmentation model will be developed taking into consideration the specific needs of part (primitives, convex components, affordance-based parts) decomposition for painting. Testing on defined benchmarks. Writing scientific report on findings of Y1.

M13-M24: The second year will be dedicated to formalize the painting trajectory generation process. We will focus on formulating the task as a deep learning problem considering it both as a visual generative task (predict a line on a surface) and as an imitation learning problem via a painting quality based reward. Assessment of work on the established benchmarks. Writing scientific report on findings of Y2.

M25- M36. The final research period will be dedicated to extending the models obtained in Y2 across domains, considering synthetic to real variation of the data, but also the introduction of new objects and painting modalities. Assessment of work on the established benchmarks. Writing scientific report on overall findings of the project. Writing of PhD thesis.
Expected target publications: It is expected that the scientific results of this thesis will be reported in the top conference in the field of robotics (ICRA, IROS, RSS) and computer vision (IEEE CVPR, IEEE ICCV, ECCV). At least one journal publication is expected on one of the following international journals: IEEE PAMI, IJCV, CVIU, RAL.
Current funded projects of the proposer related to the proposal:
Possibly involved industries/companies:Efort Europe S.r.l.

Title: Convex relaxation based methods for gray-box model identification
Proposer: Diego Regruto (POLITO), Guillaume Mercère (University of Poitiers)
Group website: https://www.dauin.polito.it/research/research_groups/sic_system_i...
Summary of the proposal: Generically speaking, data driven modeling or system identification consists in accurately estimate the parameters of dynamical models (used, e.g., to mimic the behavior of complex systems) from experimentally collected data and some prior information. Model structures encountered in system identification are often divided into three classes: white (model structure entirely based on physical equations), black (mathematical model structure with no relation with the physical equations) and gray-box models. In this project, a specific attention is paid to gray-box models, i.e., models whose structure is partially governed by some prior physical knowledge and/or first principles. Accurately estimating the parameters of a gray-box linear time-invariant state-space representation is a challenging problem especially if the number of unknowns exceeds ten, due to the fact that the model equations typically depend nonlinearly on physical parameters to be estimated. Standard nonlinear (local) optimization-based procedures often fail because the initial guesses are not in the domain of attraction of the user-defined cost function global minimum. The main goal of this project is to overcome such issues by combining linear algebra and convex-relaxation based set membership solutions which have proved their efficiency in several applied data driven modeling problems in the recent years. The activity will be developed in the context of a collaboration between the System Identification and Control (SIC) group of the DAUIN, with the Laboratory of Computer Science and Automatic Control for Systems (LCSAC) of Poitiers University led by Prof. Guillaume Mercère.
Rsearch objectives and methods: The research activities of the System Identification and Control (SIC) group in Politecnico di Torino has been mainly focused during the last decade on the development of convex-relaxation based algorithms for solving classes of polynomial optimization problems specifically arising in the framework of set-membership estimation. In parallel, data driven modeling solutions dedicated to gray box model identification have been recently developed in the Laboratory of Computer Science and Automatic Control for Systems (LCSAC) of Poitiers University. The proposed research project is focused on combining the theoretical results and numerical algorithms developed by both research groups for proposing novel effective approaches for gray-box identification able to overcome limitation of the methods already available in the literature.

The main topics of the research project are summarized in the following three parts.

Part I: set-membership approach for gray box state space model identification

Convex relaxation techniques recently developed by the SIC group for the problem of linear and nonlinear black-box model identification will be modified and extended in order to deal with the difficulties arising from the specific nonlinear character of the gray-box estimation problem.

Part II: combining linear algebra and convex relaxation based algorithms for gray box state space model identification

Recently, a new gray box model identification technique using linear algebra and closed form solutions has been suggested by LCSAC researchers. Such a technique requires specific rank constraints to be satisfied in order to guarantee that a linear algebra based solution exists and can be determined easily. When such rank constraints are not fulfilled, the linear algebra based solution boils down to a difficult nonconvex optimization problem Suitable convex relaxations approaches will be derived in order to obtain accurate approximations of the global optimal solution.

Part III: computation of model uncertainty for estimated gray-box state-space models

The main goal of this part is to propose a two-stage approach for quantifying the uncertainty of a gray-box model identified by using the techniques developed in Part I and II. In the first step parameter uncertainty intervals will be computed for a black box description of the system by means of standard set-membership techniques. Then, convex relaxation techniques will be used for converting the black-box parametric uncertainty into equivalent uncertainty intervals for the gray-box model parameters.
Outline of work plan: FIRST YEAR

January 1st – June 30th:
the first six months of the project will be devoted to the study of the literature with reference to the subject of gray box model identification, linear algebra, set-membership estimation/identification, convex relaxations of polynomial optimization problems.

Milestone 1:
Report of the results available in the literature; theoretical formulation of problem and analysis of the main difficulties/critical aspect/open problems. Results obtained from this theoretical analysis of the problem will be the subject of a first contribution to be submitted to an international conference.

July 1st – December 31st:
the second part of the first years will be devoted to the problem of extending the convex relaxation techniques previously developed by the SIC group to the case of gray-box identification.

Milestone 2:
Derivation of original convex-relaxation based algorithms for gray-box identification. Results obtained in this stage of the project are expected to be the core of a paper to be submitted to an International Journal

SECOND YEAR

January 1st – June 30th :
the first half of the second year of the project will be focused on the formulation of suitable convex relaxations for the nonlinear optimization problems arising from the linear algebra approach to gray-box identification and the derivation of suitable numerical algorithms.

July 1st – December 31st:
the objective of this part will be to derive a suitable convex-relaxation based formulation of the problem of computing the uncertainty of the parameter of a gray-box model starting from the parameter uncertainty intervals of an input-output equivalent black-box model.

Milestone 3:
Critical comparison of the different algorithms derived in the first two years of research. Results obtained in this stage of the project are expected to be the core of both a conference contribution and a paper to be submitted to an International Journal.

THIRD YEAR

January 1st – December 31st:
the last year of the project will be devoted to apply the derived gray-box identification algorithms to real-world problems in the automotive field in the context of a collaboration with FCA/CRF group.

Milestone 4:
Application of the developed methods and algorithms to experimental data taken from real-world automotive problems.
Expected target publications: Journals:
IEEE Transactions on Automatic Control, Automatica, IEEE Transactions on Control System Technologies, System and Control Letters, International Journal of Robust and Nonlinear Control.

Conferences:
IEEE Conference on Decision and Control, American Control Conference, IFAC Symposium on System Identification, IFAC World Congress
Current funded projects of the proposer related to the proposal:
Possibly involved industries/companies:Fiat Chrysler Automobile (FAC) and Fiat Research Center (CRF), which have collaborated with the System Identification and Control (SIC) group continuously in the last decade, will be involved in the application of the derived algorithms to different modeling, identification and control problems arising in the automotive fields.

Title: Techniques for rapid deployment of dependable automotive computing architectures
Proposer: Massimo Violante
Group website: http://www.cad.polito.it
Summary of the proposal: With the growing adoption of computing architectures in safety-critical applications, automotive designers face the issue of balancing cost while achieving adequate levels of risk avoidance in timely manners. Some partial solutions to this problem exist, which are essentially based on microprocessor components that are qualified for safety-critical use. The main issue is that such components are either targeting the highest risk-avoidance level (e.g., using the terminology of the automotive ISO26262 standard, ASIL D) or the lowest levels (e.g., ASIL A/B), entailing very different approaches and technological solutions that does not offer the possibility of scaling easily up/down, based on the specific needs of a given application scenario. As a result, designers must devise ad-hoc solutions that shall be validated through time-consuming activities each time a new architecture is developed.
This research proposal aims at defining a methodology and supporting reference architectures to enable rapid deployment of dependable computing architectures. The main innovation is to enable rapid deployment of dependable computing systems through pre-validated robust architecture, which can be used as off-the-shelf component without the need for re-validation. A baseline architecture is defined, and then methods to scale up/down the baseline are defined that, by construction, allow reaching specific risk-avoidance levels.
Rsearch objectives and methods: With the growing adoption of computing architectures in safety-critical applications, automotive designers face the issue of balancing cost while achieving adequate levels of risk avoidance in timely manners. Some partial solutions to this problem exist, which are essentially based on microprocessor components that are qualified for safety-critical use. The main issue is that such components are either targeting the highest risk-avoidance level (e.g., using the terminology of the automotive ISO26262 standard, ASIL D) or the lowest levels (e.g., ASIL A/B), entailing very different approaches and technological solutions that does not offer the possibility of scaling easily up/down, based on the specific needs of a given application scenario. As a result, designers must devise ad-hoc solutions that shall be validated through time-consuming activities each time a new architecture is developed.
The objective of this proposal is to provide the designers of embedded systems for automotive applications an approach to quickly deploy dependable computing architecture matching the specific needs of an application scenario.
A silver bullet solution cannot be identified as application scenarios exist in the automotive domain that are characterized by very different levels of risk avoidance, which asks for specific solutions that minimize costs while attaining the required dependability level.
In this research we propose to identify a baseline architecture that serves as a template for building a pre-validated dependable computing architecture targeting a specific risk avoidance level. The hardware/software components of the template architecture will be defined, and the validation methodology will be provided, which enable guaranteeing that any given instance of the template is a computing architecture that matches the chosen risk avoidance level.
Transformations will be defined intended for scaling the template architecture risk avoidance level, either towards higher targets by increasing cost/complexity of the computing system, or towards lower targets by simplifying the computing system.
Outline of work plan: The workplan is organized as follows:
Year 1
The first year will be devoted to reviewing the state of the art, consisting in scientific literature, as industrial best practice as documented in relevant standards such as ISO26262, and ISO25119. Moreover, interviews with industry partners will be performed to establish the requirements coming from real application scenarios. These activities are expected to last about 6 months at the beginning of the first year of activity.
The second half of the first year will be devoted to identifying the characteristics in terms of hardware/software components that shall be included in a reference computing architecture targeting a chosen risk avoidance level, and the validation methodologies that shall be put in place to sign-off the architecture as compliant with the target risk avoidance level. The result will become a template architecture that can be instantiated using available components. The main idea is to devise a ASIL-B ready architecture (most likely powered by a microcontroller and independent watchdog, including an OSEK real-time operating system with memory and time partitioning).
Year 2
The first half of the second year will be devoted for instantiating an implementation of the proposed template using a reference application scenario, possibly defined in combination with industrial partners. We will then evaluate the capability of the proposed approach to reach the target risk avoidance level the application scenario requires.
The second half of the year will then be devoted to adjust the template on the basis of the evidence collected during the experimental validation, and to identify the transformation operations to scale up/down the template risk-avoidance capabilities.
Year 3
The proposed template and scaling transformation will be used to develop two new application scenarios to evaluate the effectiveness of the proposed approach.
Expected target publications: Conferences:
- IEEE European test Symposium;
- IEEE On-line Testing Symposium;
- IEEE International Symposium on Defect and Fault Tolerance in VLSI and Nanotechnology Systems;
- IEEE International Conference on Dependable Systems and Networks;
- IEEE Design Automation and Test in Europe.
Journals:
- IEEE Transaction on Computers;
- IEEE Transactions on Industrial Electronics;
- IEEE Transactions on Dependable and Secure Computing;
- IEEE Transactions on Emerging Topics in Computing;
- IEEE Access.
Current funded projects of the proposer related to the proposal: HiEfficient ECSEL project
Possibly involved industries/companies:Huawei, ELDOR

Title: When the cloud meets the edge: Reusing your available resources with opportunistic datacenters
Proposer: Fulvio Risso
Group website: http://netgroup.polito.it
Summary of the proposal: Cloud-native technologies are increasingly deployed at the edge of the network, usually through tiny datacenters made by a few servers that maintain the main characteristics (powerful CPUs, high-speed network) of the well-known cloud datacenters.
However, we can notice that in most domestic environments or enterprises a huge number of traditional computing/storage devices are available, such as desktop/laptop computers, embedded devices and more, which run mostly underutilized.
This project proposes to aggregate the above available hardware into an “opportunistic” datacenter, hence replacing the current micro-datacenters at the edge of the network and the consequent potential savings in energy and CAPEX. This would transform all the current computing hosts into datacenter nodes, including the operating system software.
Furthermore, in order to further leverage the above infrastructure, this project would consider evolving the current software into a cloud-native approach, in which a computing device can borrow resources from a neighbor node, with potential additional advantages.
The current PhD proposal aims at investigating the problem that may arise in the above scenario, such as defining a set of algorithms that allow to orchestrate jobs on an “opportunistic” datacenter, as well as a proof-of-concept showing the above system in action.
Rsearch objectives and methods: The objectives of the present research are the following.
- Evaluate the potential impact (in terms of hardware expenditure, i.e., CAPEX, and energy savings, i.e., OPEX) of such a scenario, in order to validate the economic sustainability and the impact in terms of energy consumption.
- Extend existing operating systems (e.g., Linux) with lightweight distributed processing/storage capabilities, in order to allow current devices to host “foreign” applications (in case of availability of resources), or to borrow resources in another machines and delegate the execution of some of its task to the remote device.
- Define the algorithms for job orchestration on the “opportunistic” datacenter, which may differ considerably from the traditional orchestration algorithms (limited network bandwidth between nodes; highly different node capabilities in terms of CPU/RAM/etc; reliability considerations; necessity to leave free resources to the desktop owner, etc).
- Define a strategy to transform traditional applications (e.g., desktop applications) into cloud-native software, which can work on cloud-native infrastructure as well, along the lines of the “Kubernetes on Desktop” project , while apparently maintaining their “desktop” look and feel. This would be required for running the current applications on the cloud-native platform defined above.

Finally, a use case will be defined in order to validate the above finding in more realistic conditions. Among the possible choices, a University lab with many desktop computers, including possible “friendly users” among the students who contribute with their laptops.
Outline of work plan: The proposed research plan, which covers a subset of the possible objectives listed in the previous section, is structured as follows (in months):
- [1-10] Economic and energy impact of opportunistic datacenters
o Real-world measurements in different environment conditions (e.g., University lab; domestic environment; factory) about computing characteristics and energy consumption
o Creation of a model to assess potential savings (economic/energy)
o Paper writing
- [11-26] Job orchestration on opportunistic datacenters
o State of the art of job scheduling on distributed infrastructures (e.g., including edge computing)
o Real-world measurements of the features required for distributed orchestration algorithms (CPU/memory/storage consumption; device availability; network characteristics)
o Definition of a scheduling model that achieves the foreseen objectives, evaluated with simulations
o Paper writing
- [27-34] Experimenting with opportunistic datacenters
o Proof of concept of the defined orchestration algorithm on real platforms
o Real-world measurements of the behavior of the above algorithm in a single use-case (e.g., University lab)
o Paper writing
- [30-32] (in parallel) Transforming traditional applications into cloud-native services
o This task aims at addressing the problem of using realistic applications in the real-world validation scenario presented above. This task is delayed in the hope that the technological evolution will bring to life a new set of cloud-native desktop applications, hence making this task un-necessary.
- [35-36] Writing PhD dissertation.
Expected target publications: Top conferences:
- USENIX Symposium on Operating Systems Design and Implementation (OSDI)
- USENIX Symposium on Networked Systems Design and Implementation (NSDI)
- International Conference on Computer Communications (INFOCOM)
- ACM European Conference on Computer Systems (EuroSys)
- ACM Symposium on Principles of Distributed Computing (PODC)
- ACM Symposium on Operating Systems Principles (SOSP)

Journals:
- IEEE/ACM Transactions on Networking - IEEE Transactions on Computers - ACM Transactions on Computer Systems (TOCS)

Magazines:
- IEEE Computer
Current funded projects of the proposer related to the proposal:
Possibly involved industries/companies:

Title: Liquid computing: orchestrating the computing continuum
Proposer: Fulvio Risso
Group website: http://netgroup.polito.it
Summary of the proposal: Recent technologies such as edge and fog computing are bringing the advantages of cloud-native technologies at the edge of the network. In the near feature we expect to include also customer devices (laptop, smartphones, IoT devices) in this computing continuum, creating the so called “liquid computing” in which computing tasks and data can be executed almost anywhere.
On the other side, this scenario poses non-trivial challenges. First, the choice of the optimal location for each computing task and data source (which can change rapidly over time, particularly in case of mobile devices). Second, the management of administrative policies, as computing and storage resources can be under the control of different administrative organizations (e.g., production factory, telco edge POP, cloud data center), each one with different goals and optimization functions. In addition, also customers (e.g., Over-The-Top operators), which represent mainly resource consumers, have their objectives which may collide with the resource providers. However, it is well understood that the “liquid computing” scenario will happen only if a “win-win” solution will be defined that satisfies all the involved actors.
Given the complexity of this scenario, the current PhD is oriented to investigate the above problem focusing on the case of multi-administrative actors, in particular (1) when telecommunication operators (telcos) are active part in this game and (2) when economic considerations are included in this picture.
Rsearch objectives and methods: The candidate will pursue one (or more) of the following objectives, while guaranteeing the “win-win” property mentioned above and the necessity to establish economic relationships among parties:
- Define novel architectural paradigms, scalable algorithms, and protocols to advertise, negotiate and acquire resources in another administrative domain, with the guarantee that advertised resources and negotiated price would be shared only between the involved parties.
- Define scalable algorithms and protocols to monitor the computing/networking infrastructure, in order to evaluate, in real-time, the actual convenience of offloading a task in a foreign domain.
- Define scalable resource sharing algorithms and protocols that enable a hosting domain to keep control of its own resources (even if offered to a foreign actor), while allowing the customer to monitor the state of the remote jobs such as they were actually running in its own administrative domain.
- Define effective and dynamic network virtualization technologies that enable to extend a given administrative domain to include “foreign” resources, which are all perceived such as local.
- Define a real-world use case that will allow to validate the above finding in more realistic conditions. Among the possible choices, OTT delivering services across different domains (e.g., cloud datacenter, edge telco cloud, local infrastructure such as stadium, cloud gaming), enterprise IT services (spanning across the enterprise network but including also telco and other third parties infrastructures), and more.
Outline of work plan: The proposed research plan, which covers a subset of the possible objectives listed in the previous section, is structured as follows (in months):
- [1-5] State of the art
o Scheduling and job allocation
o Mathematical foundation: game theory / auctions
- [6] Paper writing: survey and state-of-the-art of distributed task scheduling
- [7] Extension of a cloud toolkit to include also end-devices (e.g., through Rancher K3s).
- [8-10] Design and implementation of a scheduler for intra-domain jobs, including quality-related parameters such as latency, bandwidth, expected response time.
- [11-14] Validation through simulations (for large-scale testbeds) and by physical setup, with the collection of real metrics.
- [15] Paper writing (conference).
- [16-21] The multidomain case: extension of the scheduling algorithm for a multidomain scenario, in which administrative policies (in additional to physical constraints) play a fundamental role.
- [22-24] Validation through simulations (for large-scale testbeds) and by physical setup, with the collection of real metrics.
- [25-26] Paper writing (journal).
- [25-36] In parallel, start a low-intensity task oriented to create and guide a working group in an existing open-source cloud orchestration toolkit (e.g., Kubernetes) in order to create a standard for multi-domain interconnections (protocols, API), mimicking what the Border Gateway Protocol (BGP) does in case of multi-domain networks. Possibly this should generate a publication focuses on the defined protocol and programming interface.
- [27-31] Large scale setup and monitoring of run-time data with the help of a use-case partner (possible choices: (1) factory, (2) hospital).
- [31-34] Possible evolutions of the algorithms; push of the solution toward open-source software; paper writing (magazine).
- [35-36] Writing PhD dissertation.
Expected target publications: Top conferences:
- USENIX Symposium on Operating Systems Design and Implementation (OSDI)
- USENIX Symposium on Networked Systems Design and Implementation (NSDI)
- International Conference on Computer Communications (INFOCOM)
- ACM European Conference on Computer Systems (EuroSys)
- ACM Symposium on Principles of Distributed Computing (PODC)
- ACM Symposium on Operating Systems Principles (SOSP)

Journals:
- IEEE/ACM Transactions on Networking
- IEEE Transactions on Computers
- ACM Transactions on Computer Systems (TOCS)

Magazines:
- IEEE Computer
Current funded projects of the proposer related to the proposal:
Possibly involved industries/companies:Telecom Italia Mobile, TOPIX (Torino Piemonte Internet Exchange)

Title: Attention-guided cross domain visual geo-localization
Proposer: Barbara Caputo
Group website: https://www.dauin.polito.it/personale/scheda/(nominativo)/barbara...
Summary of the proposal: Photo geolocation is a challenging task since many photos offer only few cues about their location. For instance, an image of a beach could be taken on many coasts across the world. Even when landmarks are present there can still be ambiguity: a photo of the Rialto Bridge could be taken either at its original location in Venice, or in Las Vegas. Traditional computer vision algorithms rely on the features provided to them during training. While this can lead to some degree of success when the data distribution of training and test data are the same, the problem becomes arduous when there is a domain shift between the two distributions. The goal of this PhD is to study the problem of visual geo-localization across visual domains. We will leverage over the intrinsic spatial connotation of place images and combine attention mechanisms with modern domain adaptation algorithms, in order to obtain perceptual representations that can be used for visual place recognition, as well as for content based image retrieval, able to close the domain gap differently on different parts of the images. Experiments will be conducted on publicly available databases as well as on data collections created during the project.
Rsearch objectives and methods: The problem of assigning a geo-localization label to an image has been extensively studied in the computer vision literature. The most important challenges come from the complexity of the concepts to be recognized and from the variability of the conditions where the query images are captured [a,b]. Less researched is how to deal with shift in the distribution of the data used for training of the models with respect to those presented to the architecture at training time. The objective of this PhD is to merge together research on attention driven visual geo-localization with unsupervised domain adaptation. We will do so by leveraging over research on how to localize the spatial roots of domain shift in images [c, d]. Indeed, visual changes like those mentioned above have a clear spatial connotation, hence it is possible to localize and spatially ground the domain shift between source and target data. By doing so, we expect to be able to develop spatially localized domain adaptation architectures that will make it possible to recognize important landmarks in a scene even in presence of severe domain changes, or when a large part of a scene is occluded because of group photos. Preliminary work in this direction holds promise [e]. The algorithms developed in this PhD will be tested on publicly available databases, so to allow for a fair comparison with the current state of the art, as well as on a new database, to be acquired during the PhD, consisting of images of the most relevant Italian cities.

[a] T. Weyand etal. Planet-photo geolocalization with convolutional neural networks. Proc. ECCV 2016.
[b] M. Mancini etal. Learning deep NBNN representations for robust place categorization. IEEE Robots, and Automation Letters, 2017.
[c] T. Tommasi etal. Learning the roots of visual domain shift. Proc. ECCV16
[d] G. Angeletti, etal. Adaptive deep learning through visual domain localization. Proc. ICRA18
[e] G. M. Berton etal. Adaptive-attentive geolocalization from few queries: a hybrid approach. Proc. WACV20.
Outline of work plan: The research workplan will be articulated as follows:

M1-M6: Implementation and testing of [a,b,c,d] on reference benchmark databases in conjunction with off-the shelf state of the art in visual place geo-localization; acquisition and annotation of a first version of a database containing images of Italian cities (Turin, Milan, Genoa, Rome etc) using Google Street View; implementation and testing of relevant baselines in the literature.

M7-M12: Design and implementation of a cross-domain local loss for adaptive geo-localization. Assessment of work on established benchmarks and comparison with previous work. Assessment of work on first version of the database of Italian cities. Writing of scientific report on findings of Y1.

M13-M24: Design and implementation of a cross-domain local loss for geo-localization robust to extreme domain shifts like selfies. Extension of the database of Italian cities to contain several images of selfies taken in such cities. Assessment of work on the established benchmarks and on the acquired database. Writing of scientific report on findings of Y2.

M25-M36: Implementation of final deep architecture for cross domain geo-localization incorporating the best results from Y1 and Y2, with principled reduction of the internal parameters of the network. Completion of the data acquisition for the database of Italian cities. Assessment of work on the established benchmarks and the new database. Writing of scientific report on overall findings of the project. Writing of PhD thesis.
Expected target publications: It is expected that the scientific results of the project will be reported in the top conference in the field of computer vision (IEEE CVPR, IEEE ICCV, ECCV). At least one journal publication is expected on one of the following international journals: IEEE PAMI, IJCV, CVIU.
Current funded projects of the proposer related to the proposal: CINI-VIDESEC (3Y, 1.2M €)
Possibly involved industries/companies:PCDM –DIS

Title: Unsupervised cross domain detection and retrieval from scarce data for monitoring of images in social media feeds
Proposer: Barbara Caputo
Group website: https://www.dauin.polito.it/personale/scheda/(nominativo)/barbara...
Summary of the proposal: Social media feed us every day with an unprecedented amount of visual data. Conservative estimates indicate that roughly 10^1-10^2M unique images are shared everyday on Twitter, Facebook and Instagram. Images are uploaded by various actors, from corporations to political parties, institutions, entrepreneurs and private citizens. For the sake of freedom of expression, control over their content is limited, and their vast majority is uploaded without any textual description of their content. Their sheer magnitude makes it imperative to use algorithms to monitor, catalog and in general make sense of them, finding the right balance between protecting the privacy of citizens and their right of expression, while fighting illegal and hate content. This in most cases boils down to the ability to automatically associate as many tags as possible to images, which in turns means determining which objects are present in a scene. This PhD project will develop algorithms to automatically tag images from social media feeds and classify them with respect to their content, developing algorithms for detection and content-based image retrieval able to work robustly when it is not possible to make strong hypothesis on the visual domain where the incoming test image has been acquired.
Rsearch objectives and methods: Object detection and content-based image retrieval have been largely investigated since the infancy of computer vision. Most approaches assume that training and test data come from the same visual domain. Some authors have started to investigate the more challenging scenario where the training data come from a visual source domain, and the learned algorithm is deployed at test time in a different target domain. casting the problem in the unsupervised domain adaptation framework. This approach is not suitable, neither effective, for monitoring social media feeds, where two key challenges must be considered: (1) to adapt to the target data, these algorithms need first to collect feeds, and only after enough target data has been collected they can learn to adapt and start performing on the incoming images; (2) even if the algorithms have learned to adapt on target images collected from the feed up to time t, there is no guarantee that the images that will arrive from time t+1 will come from the same target domain.

The main objective of this PhD thesis is to investigate this largely unexplored scenario, developing algorithms able to adapt from scarce, non annotated data coming from varying target domain, in the object detection and image retrieval settings. The two scenarios will be analyzed together both for their relevance in the social feed monitoring, both for the cross-fertilization that historically have brought one on the other: indeed, object detectors are used to tag images, and the content of images primes detectors offering contextual cues. We will explore the use of self-supervised approaches for adaptation, that recently have proved to be highly effective in domain adaptation tasks [a], and will explore the challenge of moving adaptation from the feature to the metric level, a necessary step for dealing effectively with retrieval over large scale training data. Preliminary work shows promising results [b].

[a] F. M. Carlucci etal. Domain generalization by solving jigsaw puzzles. Proc. IEEE CVPR 2019.
[b] A. D’Innocente etal. One-shot unsupervised cross-domain detection. Proc. ECCV 2020.
Outline of work plan: The research workplan will be articulated as follows:

M1-M4: Implementation and testing of [a] on reference benchmark databases in conjunction with off-the shelf state of the art in object detection and image retrieval; acquisition and annotation of a social medial database; implementation and testing of relevant baselines in the literature.

M5-M12: Design and implementation of a cross-domain loss for adaptive object detection from scarce data in a social media monitoring scenario. Assessment of work on the on the established benchmarks. Testing of the effectiveness of various self-supervised tasks for the detection setting. Writing of scientific report on findings of Y1.

M13-M24: Design and implementation of a cross-domain loss for adaptive image retrieval from scarce data in a social media monitoring scenario, acting at the metric level rather than at the feature level. Testing of the effectiveness of various self-supervised tasks for the retrieval setting. Assessment of work on the on the established benchmarks. Writing of scientific report on findings of Y2.

M25-M36: Implementation of final deep architecture for the monitoring of social media feeds incorporating the best results from Y1 and Y2, with principled reduction of the internal parameters of the network. Assessment of work on the established benchmarks. Writing of scientific report on overall findings of the project. Writing of PhD thesis.
Expected target publications: It is expected that the scientific results of the project will be reported in the top conference in the field of computer vision (IEEE CVPR, IEEE ICCV, ECCV). At least one journal publication is expected on one of the following international journals: IEEE PAMI, IJCV, CVIU.
Current funded projects of the proposer related to the proposal: CINI-VIDESEC (3Y, 1.2M €)
Possibly involved industries/companies:PCDM –DIS

Title: Cross Modal Neural Architecture Search
Proposer: Barbara Caputo
Group website: https://www.dauin.polito.it/personale/scheda/(nominativo)/barbara...
Summary of the proposal: Determining an optimal architecture is key to accurate deep neural networks (DNNs) with good generalisation properties. Neural architecture search (NAS) can potentially reduce the need for application-specific expert designers allowing for a wide-adoption of sophisticated networks in various industries. It has been showed that, by applying such algorithms, the resulting architectures can indeed outperform human-designed state-of-the-art convolutional networks. Researchers have used a wealth of techniques ranging from reinforcement learning, where a controller network is trained to sample promising architectures, to evolutionary algorithms that evolve a population of networks for optimal DNN design. Still, these approaches are inefficient and can be extremely computationally and/or memory intensive. Moreover, all existing approaches assume that training and test data are generated by the same underlying distribution. This is often not true, as usually the data seen by any algorithm at deployment time are generated by a different data distribution. The goal of this PhD thesis will be that of developing computationally efficient cross modal neural architecture search approaches. Although the application domain of the thesis will be visual recognition, it is expected that the results obtained in the thesis will be general and of interest for the machine learning community at large.
Rsearch objectives and methods: Neural Architecture Search (NAS) has the potential to discover paradigm-changing architectures, removing the need for a human expert in the network design process. While significant improvements have been achieved, this has taught us little about why a specific architecture is more suited for a given dataset. We attribute this to two main issues: (i) reliance on over-engineered search spaces and (ii) the inherent difficulty in analyzing complex architectures. Regarding the first issue, current NAS methods often restrict the macro-structure and search only the micro-structure at the cell level (see [a,b] and references therein), focusing on which operations to choose but fixing the global wiring pattern. This leads to high accuracy but restricts the search to local minima. The second issue appears hard to solve, as analyzing the structure of complex networks is itself a demanding task for which few tools are available.

A promising idea is that of moving the focus towards network generators, as the whole network can then be represented by a small set of parameters. This idea, first introduced by [a], offers many advantages for NAS: the smaller number of parameters is easier to optimize and easier to interpret when compared to the popular categorical, high-dimensional search spaces. Furthermore it allows the algorithm to focus on macro differences (e.g. global connectivity) rather than the micro differences arising from minor variations with little impact on the final accuracy. This idea has been recently explored in [b], with very promising results.

The goal of this PhD is investigating this NAS framework within the context of domain adaptation for visual recognition, with a particular focus on self-supervised approaches [c] that are highly effective for this task. We expect the results obtained in this thesis to be relevant for the computer vision and machine learning communities.

[a] S. Xie, etal, “Exploring randomly wired neural networks for image recognition,” arXiv:1904.01569, 2019.
[b] B. Ru etal. Neural Architecture Generator Optimization. Proc. NeurIPS 2020.
[c] F. M. Carlucci etal. Domain generalization by solving jigsaw puzzles. Proc. IEEE CVPR 2019.
Outline of work plan: The research workplan will be articulated as follows:

M1-M6: Implementation and testing of [b] integrated with [c,d] on reference benchmark databases; implementation and testing of relevant baselines in the literature.

M7-M12: Design and implementation of a NAS framework for adaptive object detection and classification computationally efficient. Assessment of work on the on the established benchmarks. Testing of the effectiveness of various self-supervised tasks for the detection and classification settings. Writing of scientific report on findings of Y1.

M13-M24: Design and implementation of an integrated self-supervised, graph based NAS architecture. Assessment of work on the on the established benchmarks. Writing of scientific report on findings of Y2.

M25-M36: Implementation of final deep architecture incorporating the best results from Y1 and Y2, with principled reduction of the internal parameters of the network. Assessment of work on the established benchmarks. Writing of scientific report on overall findings of the project. Writing of PhD thesis.
Expected target publications: It is expected that the scientific results of the project will be reported in the top conference in the field of computer vision (IEEE CVPR, IEEE ICCV, ECCV). At least one journal publication is expected on one of the following international journals: IEEE PAMI, IJCV, CVIU.
Current funded projects of the proposer related to the proposal: Residual ICT funding + internal funding PI
Possibly involved industries/companies:

Title: Generative modelling for improved decision making in data-limited applications
Proposer: Enrico Macii, Santa Di Cataldo
Group website: https://eda.polito.it/
Summary of the proposal: The unprecedented explosion of ICT and Artificial Intelligence technologies is driving a revolution that is transforming many important sectors, including industry, healthcare and finance. For example, their applications in manufacturing has led to the advent of the so-called Industry 4.0. Supervised machine learning remains the go-to approach for many practical applications in this context, with deep networks (Convolutional Neural Networks in particular) being the state-of-the-art for any tasks involving image data. Nonetheless, training these models require massive amount of annotated data, which makes them impractical in many real-world applications. Thanks to their capability of generating new samples from the same distribution of the training data, generative models have a huge potential in solving this issue. Among the others, Generative Adversarial Networks (GANs) is a novel class of deep generative models which has recently gained significant attention, as it is able to learn complex and high-dimensional distributions implicitly. However, their application to real-world sectors, with special regard to the highly regulated ones (e.g. the manufacturing or medical fields) is still at its early stages. This research proposal aims to investigate and advance the state of the art of generative modelling for improved decision making in different types of data-limited scenarios.
Rsearch objectives and methods: The main goal of this PhD proposal is the investigation of generative modelling for solving decision making problems in data-limited situations, minimising the need for accessing real data to make informed decisions. This involves two typical scenarios:
1) the training data is difficult to obtain, or it is available in limited quantity.
2) obtaining the training data is not difficult. Nonetheless, it is either difficult or economically impractical to have human experts labelling the data.

In both scenarios, generative models (and GANs in particular) can play a major role, serving as the backbone for solutions involving either data augmentation, synthetic image generation as well as semi-supervised learning strategies.

In this regard, the objectives of the PhD program are the following:
- Investigating the state of the art, challenges and potentials of different families of generative models and their use in data augmentation, synthetic data generation and semi-supervised learning applications.
- Seek original solutions to challenges related to training stability, generalization and computational problems of existing methods.
- Designing new frameworks for data augmentation, synthetic data generation and semi-supervised learning using GANs as the backbone
- Applying the designed solutions to different types of image classification tasks suffering data-limitation problems (for example, in industrial applications).
Outline of work plan: Phase 1. The candidate will be introduced to the problem of supervised Machine Learning with lack of training data, with special regard to image classification tasks. He/she will investigate the state-of-the art of data augmentation and synthetic data generation with unsupervised generative adversarial networks, with the aim of identifying and addressing computational challenges of the existing solutions.

Phase 2. The candidate will investigate the problem of semi-supervised learning with generative adversarial networks, addressing scenarios with small amount of labelled data and large amount of unlabelled data during the training.

Phase 3. The candidate will exploit and validate the proposed frameworks on established computer vision benchmarks, artificially modified to simulate situations with training data/annotation scarcity, as well as in real-world case studies affected by real data limitation issues (see related projects sections).

The above activities should not be considered as stand-alone tasks; they may overlap, indeed, in order to achieve a more efficient integration and higher quality of results.
Expected target publications: Targeted conferences:
- IEEE CVPR
- IEEE ICCV
- IEEE ECCV
- IEEE ICPR

Targeted journals:
- Pattern Recognition
- Pattern Recognition Letters
- IEEE Transactions on Image Processing
- IEEE Transactions on Pattern Analysis and Machine Intelligence

As well as conferences and journals targeting the specific case-study applications (e.g. smart manufacturing domain)
Current funded projects of the proposer related to the proposal: The proposed solutions can be applied to several case studies in the field of Smart Manufacturing and Industry 4.0, involving the design of supervised Machine Learning techniques in data-limited situations.
Example of related funded projects:
AVISPA - Automation of Visual Inspection and Finishing Processes for Aero-engines
FCA-POLITO - Quality assurance for additive manufacturing
IMPACT - IMplementazione della Produzione Additiva CompetiTiva
Possibly involved industries/companies:FCA, Avio Aero, Prima Industrie

Title: PoliTO-EURECOM PhD on methods for modelling software protection as a risk analysis process
Proposer: Cataldo Basile (POLITO), Davide Balzarotti (EURECOM)
Group website: https://security.polito.it/
Summary of the proposal: The last years have seen an increase of Man-at-the-End (MATE) attacks against software applications, both in number and severity. Indeed, the software contains sensitive assets, from cryptographic keys and intellectual property (e.g., algorithms), which need to be kept confidential, to critical functions (e.g., license check or industrial processes), which need to be protected from tampering.
MATE attackers have full control over the devices where they attack the software (white-box). They may use simulators, debuggers, disassemblers, and all kinds of static analysis, dynamic analysis, and tampering tools.
However, software protection, as a research field, is still in his infantry. Protections such as obfuscation, software and remote attestation, anti-debugging, software diversity, and anti-emulation do not aim to prevent MATE attacks completely. They aim at mitigating the risk that assets are compromised in a given time frame. Hence, they delay potential attackers and lower the expected attacks Return Of Investment (ROI).
Fuzzy concepts and techniques dominate the design and evaluation of software protections. Security-through-obscurity is omnipresent in industry, protection tools and consultancy are expensive and opaque, there is no commonly accepted method for evaluating protection effectiveness, and we lack any form of standardization.
The Ph.D. proposal has a high-level research objective: to perform research in software protection to progress towards the formulation of software protection as a Software Risk analysis process.
Rsearch objectives and methods: The main research objective is defining models for the automatic protection of software assets and their validation through empirical experiments. The ambition is to progress towards a system that helps developers with limited or no knowledge about software protection to defend their assets.
The candidate will improve the existing definitions of potency, the metric that determines the effectiveness of software protections against (human) attackers. The objective is to develop predictive models that can reliably estimate the effectiveness of protections on code fragments before applying them. To this purpose, one research objective is modelling the complexity of reverse engineering code to acquire understanding useful to mount attacks. Empirical experiments will be designed to determine the complexity of comprehension and tampering tasks and how these are affected by the presence of software protections.
Furthermore, the candidate will investigate how to estimate the stealth of software protections, the property that measures if protected assets are visible to attackers. A related research objective is improving the existing models to optimize the stealth of protections.
The candidate will investigate methods to measure the resilience, which measures the resistance of software protections against attack tasks that aim at removing them via automated tools and approaches. To this purpose, the investigation will use advanced AI methods, e.g., Automatic Exploit Generation.
We expect to build predictive models to categorize, assess, and estimate - in an objective way - threats and mitigations.
Outline of work plan: The initial phases of the Ph.D. will be devoted to the formalization of the risk analysis framework of the software protection, to contextualize all the research objectives.
The student will start from the model of the potency. The most challenging task is building predictive models that estimate the effectiveness of protections before they are applied. Measuring ex-post is a very time/resource consuming task that may easily last minutes, rendering impossible an optimized selection of the protections.
Together with protection-specific parameters, several co-factors will be considered. Objective metrics computed on the code both before and after the protection (e.g., LOCs, Halstead, cyclomatic complexity, already used by with limited results) will be complemented with metrics based on the semantics of the code to protect.
During the period at EURECOM, the candidate will focus on the definition of models of human comprehension when reverse engineering binaries. The model will consider how attackers join forces with analysis tools that produce artifacts and abstractions like CFG/DDG, traces, disassembled, and decompiled code.
At PoliTO, the student will extend the modelling phase to tampering tasks and will collaborate in the design of experiments to empirically assess the impact of protections against attack tasks. Experimental data will also work as validation of the models.
In parallel, models of stealth will be formalized by means of machine learning methods defined to identify protected assets and classify the used protections. Models of resilience will be based on the effectiveness of the Automatic Exploit Generation, a family of techniques used to automatically determine and exploit vulnerabilities in applications, which will be improved to perform the risk assessment of software assets.
The student will spend approximately 50% of his time in each institution. A 3-6 months internship is possible in a third institution to acquire competencies that may emerge as needed.
Expected target publications: We expected at least two publications on top-level cybersecurity conferences and symposia (e.g., ACM CCS, IEEE SandP, IEEE EuroSP).
Results about software protection metrics (potency, stealth, resilience) will be submitted to top-tier journals in scope (e.g., ACM Transactions on Software Engineering, ACM Transactions on Privacy and Security, IEEE Transactions on Dependable and Secure Computing, IEEE Transactions on Emerging Topics in Computing)
We expect results for at least one paper about the empirical assessments of software protections to be submitted to journals dealing with empirical methods (e.g., Empirical Software Engineering, Springer)
Current funded projects of the proposer related to the proposal:
Possibly involved industries/companies:

Title: Combining Meta-Learning and Self-Supervision for Test-Time Adaptation
Proposer: Tatiana Tommasi
Group website: https://www.dauin.polito.it/personale/scheda/(nominativo)/tatiana...
Summary of the proposal: Deep Learning models are highly sensitive to uncontrolled data shifts during testing, still this condition appears in many practical applications, for instance in case of unexpected weather conditions or visual sensor degradation. Domain adaptation approaches have been developed to mitigate this problem, but the setting becomes extremely challenging also for them when only one test sample is available and we need to both adapt and predict on it. This is the typical scenario encountered when monitoring social media feeds: even if some algorithm has learned to adapt on target images from the feeds up to time t, there is no guarantee that the images that will arrive from time t+1 will come from the same target domain. The goal of this work will be to develop test-time adaptive learning models and show how they can support object recognition and detection in cross-domain visual tasks.
Rsearch objectives and methods: One possible solution to the described challenges can be obtained by exploiting the power of self-supervised learning [1,2,3]. Since it does not need any manual annotation, the self-supervised task (e.g. rotation recognition) can be added as an auxiliary objective in the network and runs seamlessly on each single test image, supporting adaptation to the style of every new instance.
This strategy shares its aim of learning from limited input information with another family of techniques, known as meta-learning [4,5]. It can be broadly defined as a class of machine learning models that learn how to learn: the network is guided to distill knowledge from few-shot meta-tasks during training, thus it is inherently prepared for the scenario that will be faced at test time. Meta-learning-based methods have already demonstrated their effectiveness in providing better data efficiency, improved ability in exploiting knowledge transfer, as well as in enhancing robustness and generalization of the training process.
This project is dedicated to combining for the first time self-supervised and meta-learning to tackle test-time cross-domain adaptation. Besides considering the object classification task for an initial analysis, we will focus also on object detection, where meta-learning can provide a significant support to single-sample test time adaptation, but integrating it with the localization and recognition objectives is more challenging.

[1] “Self-supervised Visual Feature Learning with Deep Neural Networks: A Survey”, TPAMI 2020
[2] “One-Shot Unsupervised Cross-Domain Detection”, ECCV 2020
[3] “On the Effectiveness of Image Rotation for Open Set Domain Adaptation”, ECCV 2020
[4] “Meta-Learning in Neural Networks: A Survey”, Preprint Arxiv April 2020
[5] “Meta-Learning with Warped Gradient Descent”, ICLR 2020
Outline of work plan: M1-M6: The initial step consists in an extensive study of the most recent cross-domain classification and detection models. The purpose is choosing the datasets (e.g. images from social media monitoring, autonomous driving) and the evaluation metrics, setting the baselines, and analyzing the limits of the most recent existing models in the new test-time adaptive scenario. This preliminary study should also confirm the effectiveness of the auxiliary self-supervised task as discussed in [2].

M6-M12: A second step consists in coding the logic of meta-learning within the self-supervised task for cross-domain object classification. Differently from previous use of meta-learning [4], our goal is to apply it on the auxiliary task to support adaptation, rather than on the main classification task. In particular we plan to exploit gradient-optimization based meta-learning strategies and assess their performance on test-time object classification. Assessment of work on the established benchmarks. Writing scientific report on findings of Y1.

M13-M24: The second year will be dedicated to extending the defined model to the detection task. We will study solutions based both on gradient-optimization and pre-conditioning matrices [5]. Moreover, we plan to investigate the introduction of a memory module in the learning process: although each target sample may be different from the previous one, storing their information can be helpful for long term future adaptation. Assessment of work on the established benchmarks. Writing scientific report on findings of Y2.

M25-36 Finally we will extend the defined model also to open set and incremental class learning, supposing that a new sample observed ad deployment time might not only come from a different domain but also contain a new class. In these cases the model should first of all reject the prediction on the basis of a low class confidence [3], and in a second learning step the new class could be integrated in the recognition model. Assessment of work on the established benchmarks. Writing scientific report on overall findings of the project. Writing of PhD thesis.
Expected target publications: It is expected that the scientific results of this work will be reported in the top conference in the field of computer vision (IEEE CVPR, IEEE ICCV, ECCV) and robotics (ICRA, IROS, RSS). At least one journal publication is expected on one of the following international journals: IEEE PAMI, IJCV, CVIU, RAL.
Current funded projects of the proposer related to the proposal:
Possibly involved industries/companies:

Title: Autonomous systems' reliability and security
Proposer: Ernesto Sanchez
Group website: http://www.cad.polito.it/
Summary of the proposal: Autonomous systems (A.S.) have been a subject of great interest during the last years. In fact, autonomous devices have been investigated and proposed by very different actors: from the automotive industry to hospital UV disinfection robots; from automatic stacking systems to unmanned aerial vehicles. In order to support the software complexity, A.S. may include some microprocessor cores, different memory cores, and hardware accelerators in charge to perform the Deep Artificial Neural Network computations.
An emerging problem is the verification, testing, reliability, and security of Autonomous Systems, and in particular, regarding the computational elements involved in the artificial intelligence computations.
These very complex systems are still lacking a holistic analysis and comprehension related to reliability and security. It is for example unclear how to functionally verify the behavior of a DNN, or what may happen when the autonomous system is affected by a fault from both points of view: reliability and security.
During this project, the Ph.D. candidate will study from the hardware and software perspective, how to improve the reliability and security of autonomous systems based on AI solutions.
Rsearch objectives and methods: The Ph.D. proposal aims at studying the current design, verification and testing methodologies that try to guarantee a correct implementation of AI based systems in A.S. with particular interest on the available solutions to increase the systems reliability and security. During the initial phase, a set of benchmarks that will provide the suitable cases of study for the following research steps are defined. Two main types of AI systems in A.S. will be analyzed: the first one is based on hardware accelerators that exploit for example FPGA implementations; and the second one implements the A.I. algorithm supported by components-off-the-shelfs (COTS) such as systems that embeds a high performance processor core.
Regarding the testing issues, it is necessary to define a set of fault models supported by a failure analysis of hardware implementations of the different systems and in particular to the ones related to the Neural Network implementations.
From the reliability point of view, there is a lack of metrics able to correctly assess how reliable is an AI-based system, in fact, a study and proposal of appropriate metrics is also required at this point.
A first attempt to consider the system security of AI algorithms running on embedded systems is also performed. Regarding security, the lack of metrics and experimental demonstration make important to fulfill this gap by providing some indications about the main security criticalities and how to mitigate them in an A.S. based on embedded systems.
Finally, mitigation strategies based on self-test, error-recovery and earlier detection mechanisms will be developed for the autonomous systems studied. The final goal is to equip the AI hardware with self-test mechanisms to detect hardware errors, and possible thread intrusions thanks to the implementation of fault-tolerance and secure oriented mechanisms for increasing the reliability and security of the AI algorithm while maintaining the system accuracy.
Outline of work plan: Proposed work plan:

1. Study and identification of the most important works on design and verification of AI solutions used in autonomous systems.
2. Design and implementation of the cases of study resorting to hardware accelerators and COTS based on high performance processor cores.
3. Fault model definition and experimentation mainly resorting to the implemented cases of study. 4. Reliability and safety metrics definition.
5. Security analysis and metrics definition.
6. Mitigation strategies proposal.
The first two steps will left the Ph.D. candidate with the appropriate background to perform the following activities.
Steps 3 to 6 are particularly interesting form the research point of view, allowing the student to write papers and present them in the International conferences related to the research area. During these research phases, the student will have the possibility to cooperate with international companies such as NVIDIA and STMicroelectronics.
This proposal is complementary and strictly related to the one titled "Reliability and Safety of Artificial Neural Networks": both focus on the reliability of AI architectures for safety-critical embedded applications, but the two proposals deal with different underlying hardware architectures.
Expected target publications: The main conferences where the Ph.D. student will possibly publish her/his works are:
DATE: IEEE - Design, Automation and Test in Europe Conference
ETS: IEEE European Test Symposium
ITC: IEEE International Test Conference
VTS: IEEE VLSI Test Symposium
IOLTS: IEEE International Symposium on On-Line Testing and Robust System Design
MTV: IEEE International workshop on Microprocessor/SoC Test, Security and Verification
Additionally, the project research may be published in relevant international journals, such as: TCAD, TVLSI, ToC.
Current funded projects of the proposer related to the proposal:
Possibly involved industries/companies:NVIDIA

Title: Human-Centered AI for Smart Environments
Proposer: Luigi De Russis, Corno Fulvio
Group website: https://elite.polito.it
Summary of the proposal: Artificial Intelligence (AI) systems are widespread in many aspects of the society. Machine Learning, in particular, enabled the development of algorithms able to automatically learn from data without any human intervention. While this leads to many advantages in terms of more efficient decision processes and productivity, it also presents several drawbacks such as disregarding end-user perspectives and needs.

In this respect, Human-Centered AI (HCAI) emerged as a novel conceptual framework [1] for reconsidering the centrality of humans while keeping the benefit of AI systems. To do so, the framework builds on the idea that a system can contemporary exhibit high levels of automation and high levels of human control.

The Ph.D. proposal applies and extends the research on HCAI to smart environments, e.g., AI-powered environments equipped with Internet-of-Things devices. In such environments, AI systems typically tend to automate the activities that people perform; users, however, want to remain in control. This generates a conflict that could be tackled by adopting the HCAI framework.
This proposal aims at designing, developing, and evaluating concrete HCAI systems to support inhabitants of smart environments. Also, it aims at extending the understanding of the HCAI framework’s principles and providing valuable lessons for different fields.

[1] Ben Shneiderman (2020) Human-Centered Artificial Intelligence: Reliable, Safe and Trustworthy, International Journal of Human-Computer Interaction, 36:6, 495-504
Rsearch objectives and methods: The main research objective is to investigate solutions for designing and developing HCAI systems in smart environments. A particular focus will be on how the adoption of the HCAI framework can bring tangible benefits to users and to the smart environments research field, while extending the research on HCAI.

The research activities will mainly build on the following characteristics of the HCAI framework:
- High levels of human control and high levels of automation are possible: design decisions should give users a clear understanding of the AI system state and its choices, guided by human-centered concerns, e.g., the consequences and reversibility of errors. Well-designed automation preserves human control where appropriate, thus increasing performance and enabling creative improvements.
- AI systems should shift from emulating and replacing humans to empowering and “augmenting” people, as people are different from computers. Intelligent system designs that take advantage of unique computer features are more likely to increase performance. Similarly, designs that recognize the unique capabilities of humans will have advantages such as encouraging innovative use and supporting continuous improvement.

In particular, the Ph.D. research activity will focus on:
1) Study of AI algorithms and models, distributed architectures, and HCI techniques able to support the identification of suitable use cases for building effective and realistic HCAI systems.
2) Enhancement of the HCAI framework to include end-user personalization, e.g., as a way to recover from errors or to guide the system choices.
3) Development of strategies for dealing with de-skilling effects. Such effects may undermine the human skills that are needed when automation fails and the difficulty of remaining aware when some user actions become less frequent.
Such goals will require advancement both in interfaces and interaction modalities, and in AI algorithms and their integration into user-facing smart environments.
Outline of work plan: The work plan will be organized according to the following four phases, partially overlapping. Phase 1 (months 0-6): literature review about HCAI and smart environments; study and knowledge of AI algorithms and models, IoT devices and smart appliances, as well as related communication and programming practices and standards.
Phase 2 (months 6-18): based on the results of the previous phase, definitions and development of a set of use cases and interesting contexts to be adopted for building user-facing smart environments. Initial data collection for validating use cases, and possible applications of end-user personalization strategies.
Phase 3 (months 12-24): research, definition, and experimentation of HCAI systems in the defined smart environments and use cases, starting from the outcome of the previous phase. Such solutions will imply the design, implementation, and evaluation of distributed and intelligent systems, able to take into account users’ preferences, capabilities of a set of connected devices, as well as AI methods and algorithms.
Phase 4 (months 24-36): extension and possible generalization of the previous phase to include additional contexts and use cases. Evaluation in real settings of one or more of the realized systems over a significant amount of time.
Expected target publications: For each of the previously mentioned phases, at least one conference or journal publication is expected. Suitable venues might include:
ACM CHI, ACM IUI, ACM Ubicomp, IEEE Internet of Things Journal, ACM Transactions on Internet of Things, IEEE Pervasive Computing, IEEE Transactions on Human-Machine Systems, ACM Transactions on Computer-Human Interaction.
Current funded projects of the proposer related to the proposal: None at the moment.
Possibly involved industries/companies:None at the moment.

Title: Vehicular micro clouds in 5G edge/cloud infrastructures
Proposer: Claudio Casetti (POLITO), Fulvio Risso (POLITO), Jerome Harri (EURECOM)
Group website: http://netgroup.polito.it
Summary of the proposal: The topic of the joint PhD position will address the new area of vehicular micro clouds: a paradigm that aims at sharing processing capabilities, storage and networking services between cars, pedestrians, and bicyclists organized in clusters, thereby enabling cloud and edge services to extend transparently between involved actors independently of their individual capabilities. Micro-clouds are major enablers towards 5G-driven decentralized IoT architectures and distributed edge intelligence.
Rsearch objectives and methods: Starting from the identification and definition of vehicular use cases that can benefit from micro clouds (e.g., cooperative hazard detection leveraging decentralized AI/ML, localized processing of massive sensorics data, ultra-low latency consensus-based decision making, constrained device processing/storage offloading), the PhD student should focus on the study and development of solutions that integrate vehicular micro clouds and 5G edge/cloud continuum within a decentralized IoT architecture.
This activity could include as possible directions of investigation:
- Novel distributed task scheduling techniques
- Novel resource sharing models supporting multiple administrative domains
- Novel hybrid cloud architectures
- Novel resource offloading algorithms suitable in case of short-lived/unreliable connections between devices, some of which may be energy-constrained
- Load balancing between 5G V2V and V2I communication techniques in the micro cloud
- On-demand opportunistic micro-cloud discovery and formation

This topic is expected to have a notable impact on different vertical sectors (e.g., automotive, telecommunications, smart cities) thanks to (a) the ongoing trend, supported by the 5G architecture, of creating edge-based micro-cloud facilities, (b) the increasing “cloudification” of any computing/storage resource, (c) novel vehicular applications (e.g., autonomous driving).
Outline of work plan: It is expected that the PhD student will spend roughly 50% of her/his time in each institution, although plans could be altered due to lingering disruption caused by COVID-19 restrictions.
During the course of the PhD regular calls will be scheduled between the PhD student and the Tutors to have updates on the ongoing work and to co-advise the student, and the remote teaching, lab access, progress monitoring or videoconferencing technologies developed during the COVID-19 lockdown will be used to enable full access to both POLITO or EURECOM resources, regardless of the location of the PhD student.
The cooperation will also be organized at the tutor level, with mutual visits during the course of the joint PhD (at least twice at the occasion of the mid-term and final exam).
Expected target publications: Top conferences:
- USENIX Symposium on Networked Systems Design and Implementation (NSDI)
- International Conference on Computer Communications (INFOCOM)
- ACM SIGCOMM

Journals:
- IEEE/ACM Transactions on Networking
- IEEE Transactions on Computers
- ACM Transactions on Computer Systems (TOCS)
- IEEE Transactions on Cloud Computing
- IEEE Transactions on Vehicular Technology

Magazines:
- IEEE Computer
- IEEE Networks
- IEEE Communications
Current funded projects of the proposer related to the proposal:
Possibly involved industries/companies:

Title: Urban intelligence
Proposer: Silvia Chiusano
Group website: dbdmg.polito.it
Summary of the proposal: In the urban ecosystem a multitude of strongly intertwined systems coexists, varying from people sociality to transport systems. While each of these city facets already represents in itself a complex system, their interconnection is definitively a challenging scenario.

Urban intelligence entails the acquisition, integration, and analysis of big and heterogeneous data collections generated by a diversity of sources in urban spaces to profile the different facets and issues of the urban environment. It can unearth a rich spectrum of knowledge valuable to characterize citizen behavior and identify weaknesses and strengths of the services provided or even devise new ones. Thus, urban intelligence plays a key role in achieving a smart and sustainable city.

However, data analytics on urban data collections is still a daunting task, because they are generally too big and heterogeneous to be processed through machine learning techniques currently available. Thus, today's urban data give rise to a lot of challenges that constitute a new inter-disciplinary field of data science research.
Rsearch objectives and methods: The PhD student will work on the study, design and development of proper data models and novel solutions and for the acquisition, integration, storage, management and analysis of big volumes of heterogeneous urban data.

The research activity involves multidisciplinary knowledge and skills including database, machine learning techniques, and advanced programming. Different case studies in urban scenarios such as urban mobility, citizen-centric contexts, and healthy city will be considered to conduct the research activity.

The objectives of the research activity consist in identifying the peculiar characteristics and challenges of each considered application domain and devise novel solutions for the management and analysis of urban data for each domain. More urban scenarios will be considered with the aim of exploring the different facets of urban data and evaluating how the proposed solutions perform on different data collections.

More in detail, the following challenges in the context of data analytics will be addressed during the PhD, such as:

- Suitable data fusion techniques and data representation paradigms should be devised to integrate the heterogeneous collected data into a unified representation describing all facets of the targeted domain. For example, urban data are usually acquired by means of different sensor networks deployed in the city. Since these data are often collected with different spatial and temporal granularities, suitable data fusion techniques should be devised to support the data integration phase, and provide a spatio-temporal alignment of collected data.

- Adoption of proper data models. The storage of heterogeneous urban data collections requires the use of alternative data representations to the relational model such as NoSQL databases (e.g., MongoDB), also able to manage geo-referenced data.

- Design and development of algorithms for big data analytics. Urban data is usually charaterized by spatio-temporal coordinates describing when and where data has been acquired. Spatio-temporal data has unique properties, consisting of spatial distance, spatial hierarchy, temporal smoothness, period and trend, which entails the design of suitable data analytics methods. Moreover, huge volume of data demands the definition of novel data analytics strategies also exploiting recent analysis paradigms and cloud based platforms.

- Proper strategies will be also devised for data and knowledge visualization.
Outline of work plan: 1st Year. The PhD student will review the recent literature on urban computing to identify the up-to-date research directions and the most relevant open issues in the urban scenario. Based on the outcome of this preliminary explorative analysis, an application domain, such as urban mobility or urban air pollution, will be selected as a first reference case study. This domain will be investigated to (i) identify the open research issues, (ii) identify the most relevant data analysis perspectives for gaining useful insights, and (iii) assess of main data analysis issues. The student will perform an exploratory evaluation of state-of-the-art technologies and methods on the considered domain, and will present a preliminary proposal for the optimization techniques of these approaches.

2nd and 3rd Year. Based on the results of the 1st year activity, the PhD student will design and develop a suitable data analysis framework including innovative analytics solutions to efficiently extract useful knowledge in the considered domain, aimed at overcoming weaknesses of state-of-the-art methods.

Moreover, during the 2nd and 3rd year, the student will progressively consider a larger spectrum of application domains in the urban scenario. The student will evalute if and how his/her proposed solutions can be applied to the new considered domains as well as he/she will propose novel analytics solutions.

During the PhD, the student will have the opportunity to cooperate in the development of solutions applied to the research projects on smart cities. The student will also complete his/her background by attending relevant courses. The student will participate to conferences presenting the results of his/her research activity.
Expected target publications: Any of the following journals
IEEE TKDE (Trans. on Knowledge and Data Engineering)
ACM TIST (Trans. on Intelligent Systems and Technology)
Journal of Big Data (Springer)
Expert Systems With Applications (Elsevier)
Information Sciences (Elsevier)

IEEE/ACM International Conferences
Current funded projects of the proposer related to the proposal: CANP (La Casa nel Parco): Regione Piemonte, Piattaforma Tecnologica “Salute e Benessere”
S[m2]ART (Smart Metro Quadro): Bando Smart City and Communities; Ente finanziatore: MIUR (Ministero dell’istruzione, Università e Ricerca)
Possibly involved industries/companies:

Title: Explainable AI (XAI)
Proposer: Elena Baralis
Group website: http://dbdmg.polito.it/
Summary of the proposal: Machine learning models are increasingly adopted to assist human experts in decision making. Especially in critical tasks, understanding the reasons behind model predictions is essential for trusting the model itself. Investigating model behavior can provide actionable insights. For example, experts can detect model wrong behaviors and actively work on model debugging and improvement. Unfortunately, most high performance models lack interpretability.

The main goal of this research activity is the study of methods to allow human-in-the-loop inspection of classifier reasons behind predictions. Explanations can help data scientists and domain experts to understand and interactively investigate individual decisions made by black box models.
Rsearch objectives and methods: Exploring and understanding the motivations behind black-box model predictions is becoming essential in many different applications. Different techniques are usually needed to account for different data types (e.g., images, structured data, time series).

The research activity will consider industrial domains (e.g., the spatial domain) in which the availability of understandable explanations is relevant. The explanation algorithms will target both structured data and time series. The following different facets of XAI (Explainable AI) will be addressed.

Model understanding. The research work will address local analysis of individual predictions. These techniques will allow the inspection of the local behavior of different classifiers and the analysis of the knowledge different classifiers are exploiting for their prediction. The final aim is supporting human-in-the-loop inspection of the reasons behind model predictions.

Model trust. Insights into how machine learning models arrive at their decision allow evaluating if the model may be trusted. Methods to evaluate the reliability of different models will be proposed. In case of negative outcomes, techniques to suggest enhancements of the model to cope with wrong behaviors and improve the trustworthiness of the model will be studied.

Model debugging and improvement. The evaluation of classification models generally focuses on their overall performance, which is estimated over all the available test data. An interesting research line is the exploration of differences in the model behavior, which may characterize different data subsets, thus allowing the identification of potential sources of bias in the data.
Outline of work plan: PHASE I (1st year): state-of-the-art survey for algorithms and for XAI both for time series and structured data, performance analysis and preliminary proposals of improvements over state-of-the-art algorithms, exploratory analysis of novel, creative solutions for XAI; assessment of main explanation issues in 1-2 specific industrial case studies.
PHASE II (2nd year): new algorithm design and development, experimental evaluation on a subset of application domains; deployment of algorithms in selected industrial contexts.
PHASE III (3rd year): algorithms improvements, both in design and development, experimental evaluation in new application domains.
During the second-third year, the candidate will have the opportunity to spend a period of time abroad in a leading research center.
Expected target publications: Any of the following journals
IEEE TKDE (Trans. on Knowledge and Data Engineering)
ACM TKDD (Trans. on Knowledge Discovery in Data)
ACM TDS (Trans. On Data Science)
ACM TOIS (Trans. on Information Systems)
Information sciences (Elsevier)
Expert systems with Applications (Elsevier)
Machine Learning with Applications (Elsevier)
Engineering Applications of Artificial Intelligence (Elsevier)
Journal of Big Data (Springer)

IEEE/ACM International Conferences
Current funded projects of the proposer related to the proposal:
Possibly involved industries/companies:Collaboration in progress with ALTEC SpA.

Title: Quantum Machine Learning Applications
Proposer: Bartolomeo Montrucchio
Group website: http://grains.polito.it
Summary of the proposal: Quantum Computing (QC) is a quite new research field, mainly related to physics departments until recent years. In 2016, IBM introduced the first ever quantum computer, setting a milestone in this field, opening the topic to the computer science domain.
The main target of QC engineers will be the analysis and development of new algorithms, as well as of new technologies for building such quantum computers.
Along with IBM, several other companies and startups are developing their own development framework, boosting the interest in several fields.
The PhD candidate will be required to use a strong interdisciplinary approach for working on machine learning on quantum computers, since quantum mechanics must be considered together with specific computer engineering techniques like, just as examples, artificial intelligence and computer security.
Rsearch objectives and methods: Being a totally new paradigm, quantum computing is going to be a challenge for engineers, who would have not only to re-implement classical algorithms in a quantum way, but also explore uncharted paths of the new way of representing and elaborating information and its processing.
In the last three years, QC companies and research institutes have come up with different software stacks, appealing to a wide spectrum of possible users, from Machine Learning to Optimization to Material simulation.
These companies are trying to provide the programmer pseudo-standard APIs like the ones already available for conventional computers.
Analysis and possibly development of new algorithms will therefore be the final research objective of this research activity.
Since the QC landscape is in continuous evolution thanks to hardware improvements, the PhD student should be able to quickly adapt to these changes, carefully studying and comparing APIs for a certain application domain.
Therefore, the final purpose of the work will be to understand how to apply this quite new mindset in the computer engineering environment.
Industrial applications will be seen with particular attention, since industries will be the first to be involved in QC revolution.
In particular, problems like resource allocation and task scheduling could be analyzed, as well as novel Image Processing techniques with applications in the medical or manufacturing domain.
Outline of work plan: The work plan is structured in the three years of the PhD program:
1- in the first year the PhD student should improve his/her knowledge of quantum computing and technology, in particular since quantum mechanics and quantum computing are not seen in the previous curriculum; he/she should also follow in the first year most of the required courses in Politecnico. At least one or two conference papers will be submitted during the first year. The conference works will be presented by the PhD student him/herself based on the preliminary study of algorithms working on envisioned platforms.
2- In the second year the work will be both on designing and implementing new algorithms and on preparing a first work for a journal, together with another conference. Interdisciplinary aspects will be also considered. Credits for teaching will be also finalized.
3- In the third year the work will be completed with at least a publication in a selected journal summarizing the results of the algorithms implementation on platforms and technologies that will be selected as the most promising. The participation to the preparation of proposals for funded projects will be taken in consideration.
Expected target publications: The target publications will be main conferences and journals related to quantum computing, if possible. Since at the moment there are only a very few of them in the computing engineering field, the choice will be done selecting, if possible, those linked to IEEE and ACM, that already started publishing specifically on QC. It is important to note that interdisciplinary aspects will be considered as fundamental, since QC is now very useful for solving problems that can come from many research fields.
Current funded projects of the proposer related to the proposal: In the last two years a (funded) collaboration with General Motors (now Punch Torino ) has been done on these arguments.
Possibly involved industries/companies:Punch Torino

Title: Virtual and Augmented Reality: new frontiers in education and training
Proposer: Andrea Sanna
Group website: http://grains.polito.it/
Summary of the proposal: In recent years there has been huge growth in investment in Augmented Reality (AR) and Virtual Reality (VR) technologies. In 2015-16, investment grew 300% to a whopping $2.3bn. Education and training will be one of the strongest use cases for AR/VR as these technologies can help us to visualize concepts in a more interactive way. When teachers struggle to communicate complex ideas and ensure people are interpreting them correctly, we can transport students from looking at something complex to being virtually immersed in a real life example. Secondly, AR/VR eliminates the need for physical materials, which can be an expensive barrier to access. Third, AR/VR simulations can provide a unique perspective about (physical) phenomena, thus enhancing learning experiences.
This proposal aims to investigate new teaching paradigms based on AR/VR. The advent of consumer devices, such as, Microsoft Hololens, Oculus Rift, and many others, opens new challenges. Students can now take advantage of technologies and tools that can strongly improve traditional teaching methodologies. On the other hand, the design and the development of AR/VR applications is still a task for software developers and teachers are not usually able to create engaging AR/VR contents.
Rsearch objectives and methods: The idea of AR/VR being a knowledge transfer technologies is also confirmed by latest research in the area. Literature reviews indifferent application fields such as design and manufacturing, maintenance, surgery, or education have identified research gaps regarding AR/VR knowledge transfer abilities. Besides, these gaps were always related at least with one content-related technique: creation (Authoring), adaptation (Context-Awareness) or improvement (Interaction-Analysis) of augmented content.
Main goals of this Ph.D. are the design, the implementation and the validation of new AR/VR based tools able to change and improve the traditional teaching paradigms. AR and VR technologies can help students to get a more realistic 3D space comprehension, thus enhancing learning in several disciplines such as geometry, physics, chemistry, mechanics, architecture, medicine and so on.
Several issues have to be addressed in order to design and implement engaging and effective AR/VR applications for teaching purposes. First of all, the content creation has to be addressed; therefore, a framework to develop AR/VR tools without any particular coding skill has to be developed. This framework has to enable teachers to create their own applications; despite existing solutions, teachers have to be able to develop their own contents according to a "narrative". The definition of effective storytelling plays a key role in order to design engaging AR/VR applications according to edutainment strategies: concepts of "reward", "gratification", "levels of difficulty", etc. have to be introduced to involve students/trainees in a virtuous learning mechanism.
Moreover, a mechanism to maximize (and if it is possible for measuring) the transfer of knowledge has to be introduced. This is one of the main problems of existing AR/VR applications designed for education.
There are three main categories of knowledge: explicit, implicit and tacit. Basically, explicit knowledge is the kind that's easily written down or verbalized. Implicit learning is the application of detailed knowledge. Tacit knowledge is knowledge gained through unique experiences that are difficult to explain—having a 'feeling' for performing a task. Explicit knowledge is the most straightforward to store and transfer over, whereas implicit and tacit knowledge are trickier to capture.
It has been proved that AR and VR dramatically improve both the capture and dissemination of knowledge. In particular, it is possible to capture tacit knowledge in a teachable and repeatable manner. Therefore, the transfer of knowledge does also depend on the user experience (UX), which can be maximized only by user-centered design techniques. The design of the interface, the interaction paradigms and the generation of the augmented contents are just three issues to be investigated in order to maximize the UX.
Outline of work plan: The work plan will be defined year by year; for the first year of activity, it is expected the candidate will address the following points:
- Analysis of the state-of-the-art (in this phase, both AR/VR technologies and usual teaching methodologies for a large spectrum of disciplines will be investigated). The candidate could complete this analysis step by attending to specific Ph.D. courses related to teaching methodologies. Moreover, edutainment strategies have to be also considered as they can help the design of more attractive and engaging AR/VR applications.
- Objective definition (after the analysis step, the candidate has to be able to identify challenging open problems, thus defining concrete objectives for the next steps).
- Methods (concurrently to the objective definition step, the research methodology has to be also chosen; a right mix between theoretical formulation and experimental tests has to be identified).
- Expected outcomes (research papers, scientific contributions, significance, potential applications).

During the second year, the research activity will be mainly focused on the analysis of possible solutions, which is preparatory to the design and the implementation of AR/VR applications. In particular, the candidate will address the following points:
- Conceptual and SWOT analysis of possible solutions.
- Design of AR/VR tools for supporting bachelor and master students of computer animation and 3D modeling courses.
- Implementation of these new teaching paradigms based on augmented and virtual reality technologies
- Expected outcomes (research papers, scientific contributions).

During the last year, the research activity will be mainly focused on tests and result analysis. In particular, new tools will be used in the "3D Modeling Course" of the Design degree and master in the "Computer Animation" course of the Computer Science degree. Results obtained in terms of number of students that will pass the exam, average final marks, and student opinions will be compared with data collected for previous academic years, thus obtaining a clear assessment of the impact of AR/VR technologies on learning experiences. Expected outcomes are research papers and scientific contributions.
Expected target publications: IEEE Transactions on Education
ACM Transactions on Computing Education
International Journal of Educational Research (Elsevier)
IEEE Transactions on Visualization and Computer Graphics
ACM Transaction Human Computer Interaction
Current funded projects of the proposer related to the proposal: Not yet but a proposal named "Digital Learning" is currently under evaluation for the regional PRISM-E call.
Possibly involved industries/companies:IS_LM SrL, Archibuzz Srl, Synesthesia Srl, Conversa Srl

Title: Web-based distributed real-time sonic interaction
Proposer: Antonio Servetti
Group website: http://media.polito.it
Summary of the proposal: Recent integration of real-time audio functionalities in the web browsers such as Web Audio API and WebRTC has permitted new potentialities in web-based interactive and distributed audio systems.

Everyone, given an HTTP URL and a web application, can experiment sonic interaction even through a mobile device. For example, a performer can control the speakers of spectators’ smartphones during a performance, a visitor can connect with an installation and let the sound mediate the interaction with it.

However, the disparities across mobile device capabilities, e.g., audio latency, sound intensity, CPU speed, relative position, significantly hinders the quality of any performance and long preliminary device setup sessions are needed to “tune” the devices.

The proposed activity will address, both theoretically and practically, the problems of managing audio interaction in a mobile web application considering both technical issues such as device coordination and synchronization, and sonic issues related to the information conveyed by the sound, its meaning and emotional content.

The main technical aim is to provide guidance in the improvement of such interactive audio performances by means of automatic techniques that could compensate for the differences in the inhomogeneous set of smartphones, as well as to provide support for improved device coordination.
Rsearch objectives and methods: Many sonic interaction works have been recently proposed for web-based applications. However, almost all of them are designed to interact with a single user in an acoustically isolated environment, e.g., through headphones.

The main challenges, when dealing with multiple users and devices that share the same physical space and software application, are coordination and synchronization. In fact, different devices exhibit different behaviors because of their hardware and software capabilities. Even in a human scenario, i.e., an orchestra, a conductor is required to provide a means for synchronization and a tuning session is required to setup each instrument before the performance.

Starting from our ongoing work on web audio and streaming technologies, a systematic approach will be followed by analyzing and taking advantage of synchronization and coordination techniques discussed in literature for different scenarios. The final aim is to realize a framework that will improve collective sonic interaction when web applications are used.

Audio functionalities for web applications and mobile devices are relatively new and further studies are needed to improve their usability, stability and performance. To this aim, we plan to focus on technical issues in order to improve automatic setup of the devices that, as the instruments in an orchestra, should be tuned and synchronized. For example, it is essential to uniform their sound level, identify their position, know their latency and be able to synchronize them on the beats.

Such objectives will be achieved by using both theoretical and practical approaches. The resulting insight will then be validated in practical cases by analyzing the performance of the system with simulations and real experiments. In this regard, the research will be carried on in close cooperation with the Turin Music Conservatory, so as ¬to supplement our experience in sonic production and interaction.
Outline of work plan: In the first year, the PhD candidate will familiarize with the recently proposed Web Audio and WebRTC API for audio processing in the web browser, as well as the characteristics of the existing applications for sonic interaction and the artistic implications of the adoption of different technologies. This activity will address the creation of a framework that could allow multiple devices to coordinate and interact together in real.time through a web application, along with the definition of a set of practical use cases. Such activity, culminating in the implementation, analysis and comparison of different synchronization techniques in a web environment, is expected to lead to conference publications.

In the second year, building on the framework and the knowledge already present in the research group, new experiments for automatic tuning and synchronization of the devices will be developed, simulated and tested to demonstrate their performance and in particular the ability to improve the coordination of sound events and the interaction through the devices. The actual production of new sound art / sound design works will be crucial for this assessment. In this context, potential advantages of such techniques will be systematically analyzed. These results are expected to yield at least a journal publication.

In the third year, the activity will be expanded to study new sonic interaction experiences that can be built on top of the developed framework. In this context, this novel approach could unfold novel possibilities in the design of interfaces for musical expression and in the composition of multisource electro-acoustic music. Such proposals will target journal publications.

Throughout the whole PhD program, the Electronic Music Studio and School of the Music Conservatory of Turin will be involved in the research activity, specifically focusing on its practice-based aspects and the production of new interactive sound works.
Expected target publications: Possible targets for research publications (well known to the proposer) include IEEE Transactions on Multimedia, ACM Transactions on Multimedia Computing Communications and Applications, Elsevier Multimedia Tools and Applications, Computer Music Journal, Journal of the Audio Engineering Society, various international conferences (Web Audio Conference, New Interfaces for Musical Expression Conference, International Computer Music Conference, International Symposium on Computer Music Multidisciplinary Research, Audio Mostly Conference, Sound and Music Computing Conference, IEEE Intl. Conf. Multimedia and Expo, AES Conference, ACM WWW, ACM Audio Mostly, ACM SIGGRAPH).
Current funded projects of the proposer related to the proposal:
Possibly involved industries/companies:Turin Music Conservatory "G. Verdi" (prof. Andrea Agostini)

Title: Sparse optimization for system identification and machine learning
Proposer: Diego Regruto, Sophie Fosson
Group website: https://www.dauin.polito.it/it/la_ricerca/gruppi_di_ricerca/sic_s...
Summary of the proposal: System identification and machine learning are both concerned with building models from observed data. In the last years, the interconnections between the two fields are continuously increasing. One common aspect is the search for essential or parsimonious models, i.e., models that are as simple as possible, in order to avoid redundancies, over-fitting, and undesired high complexity. This research line is drawing an increasing attention for its technological applications, such as, the implementation of the identified models in cyber-physical systems, sensor networks and smartphones. From a mathematical viewpoint, parsimonious models can be built by using sparse optimization tools, i.e., by solving suitable optimization problems where the number of effective parameters is encouraged to be small. This sparsification can be obtained by suitable regularization methods. In the context of convex optimization, the l1 norm can be exploited for this purpose. Nowadays, non-convex regularization approaches are object of current research, which are practically more effective, while more difficult to analyze.
The main goal of this PhD project is to develop and analyze sparse optimization strategies to build parsimonious models in system identification and machine learning, with particular attention to the implementation in deep neural networks.
Rsearch objectives and methods: The objectives of this PhD project are both methodological and applications oriented. We summarize them as follows.

1) Development and analysis of non-convex regularization methods for sparse optimization, in linear and non-linear frameworks. Particular attention will be devoted to non-convex polynomial strategies. The proposed methods are expected to significantly outperform the state-of-the-art l1 methods.

2) Development and analysis of algorithms to solve the sparse optimization problems obtained in 1). More specifically we will focus on the theoretical comparison between the two most common approaches: (i) convex relaxation for polynomial optimization and (ii) iterative descent algorithms (e.g. gradient descent, alternating direction method of multipliers). Although both approaches have been widely adopted, a comparative analysis of the performance is still missing.

3) Implementation of the proposed methods in the context of deep neural network based identification of large-scale distributed systems. Particular attention will be devoted to the application to cyber-physical systems in both the automotive and industrial automation (autonomous vehicle platooning, localization and cooperative control of robot teams).
Outline of work plan: The research workplan is articulated as follows:

M1-M6: Study of the literature on sparse optimization and its applications on linear models. Implementation of sparse methods to linear system identification.

Milestone 1:
Report of the results available in the literature; theoretical formulation of problem and analysis of the main difficulties/critical aspect/open problems. Results obtained from this theoretical analysis of the problem will be the subject of a first contribution to be submitted to an international conference.

M7-M12: Development and analysis of novel non-convex and polynomial techniques to improve the performance of the state-of-the-art sparsification methods for linear system identification.

Milestone 2:
Results obtained in this stage of the project are expected to be the core of a paper to be submitted to an international journal.

M13-M24: Development and analysis of non-convex and polynomial techniques for sparse optimization in machine learning, with particular attention to deep neural networks. Implementation and testing of the proposed techniques.

Milestone 3:
Results obtained in this stage of the project are expected to be the core of both a conference contribution and a paper to be submitted to an international journal.

M25-M36: Implementation of the proposed techniques in the context of cyber-physical systems, with particular attention to large-scale distributed frameworks.

Milestone 4:
Application of the developed methods and algorithms to real-world problems in the fields of vehicle platooning/formation, localization and cooperative control of mobile robot teams.
Expected target publications: Journals:
IEEE Transactions on Automatic Control, Automatica, IEEE Transactions on Neural Networks and Learning Systems

Conferences:
IEEE Conference on Decision and Control (CDC), IFAC Symposium on System Identification (SYSID), IFAC World Congress, International Conference on Machine Learning (ICML), Conference on Neural Information Processing Systems (NeurIPS)
Current funded projects of the proposer related to the proposal:
Possibly involved industries/companies:

Title: ICT technologies for the automated assessment and management of neurological diseases
Proposer: Gabriella Olmo
Group website: https://www.sysbio.polito.it/
Summary of the proposal: The research activity of the Ph.D. candidate will take place in cooperation with CNR-IEIIT in the framework of the ongoing research project "ReHome": ICT solutions for the tele-rehabilitation of cognitive and motor disabilities originating from neurological pathologies. This project is funded by Regione Piemonte. In "ReHome", IEIIT-CNR cooperates with other institutions, industries, University departments and healthcare institutions (IRCSS and AUO), to develop remote motor-cognitive monitoring and rehabilitation services for people suffering from neurological pathologies.
In particular, the research activity of the Ph.D. candidate will take place into the framework of the ten-years lasting cooperation with the neurologist of the Istituto Auxologico Italiano (IRCSS) and with the Professors of the Department of Neuroscience "Levi Montalcini" of the University of Turin.
Rsearch objectives and methods: The Ph.D. research activity of the candidate aims to develop a platform for the remote monitoring and rehabilitation of patients affected by neurological diseases, in particular Parkinson's Disease and Stroke.
In this context, the candidate activity will address:
- Body tracking techniques based on non-invasive single and multi-camera systems (2D and 3D RGB-Depth cameras), possibly combined with wearable inertial measurement units.
- Motion analysis and body tracking based on multidimensional data streams for normal/pathological motion characterization.
- Kinematic analysis of the body movements of neurological subjects performing clinical tests, to achieve information related to the patient's conditions and on the effectiveness of the rehabilitation therapy, in accordance with clinical assessments scales.
Outline of work plan: During the first year, the candidate will be introduced in the research group and will begin his/her studies in the field and get accustomed with the instrumentation and the existing SW and HW tools, in particular the multi-camera systems (2D and 3D RGB-Depth cameras) and optoelectronic motion capture facilities of EHW Lab at CNR IEIIT. He/she will help refining and testing available tools for movement analysis, focused on the classification between pathological and healthy subjects.

The second and third years will be mainly devoted to research and developments of innovative ML/AI tool, and to the support to neurologists and clinicians for the testing and subsequent clinical trial, namely:

- Development of machine learning algorithms to exploit the correlation between kinematic features and clinical scores.
- Development of supervised and unsupervised classifiers for ordinal data for the automated assessment of the patient's neurological conditions.
- Exploitation of multidimensional data to perform patient's classification in accordance with the relevant clinical scales, and the evaluation of the rehabilitation effectiveness.
- Technical support to the clinicians during the experimental campaigns for the acquisition of clinical and kinematic data from neurological subjects.

The dissemination of the results will be pursued during all the Ph.D. period.
Expected target publications: We plan to have at least a journal paper published per year in Q1 journals relevant to the Computer Science and/or Biomedical Engineering domain.
Examples of target journals:
IEEE Transactions on Biomedical Engineering
IEEE Journal on Biomedical and Health Informatics
IEEE Access
IEEE Journal of Translational Engineering in Health and Medicine
Elsevier Journal "Parkinsonism and Related Disorders"
Current funded projects of the proposer related to the proposal: "ReHome" project (Regione Piemonte)
Possibly involved industries/companies:CNR-IEIIT

Title: Simulation and Modelling of Energy Demand Flexibility in energy communities
Proposer: Enrico Macii, Lorenzo Bottaccioli, Edoardo Patti
Group website: http://www.eda.polito.it
Summary of the proposal: The need for demand flexibility as a tool for balancing the grid is becoming necessary from distribution to transmission networks to cope with the increasing penetration in the electricity mix of distributed energy sources, especially variable renewable power sources such as wind and photovoltaic. Demand flexibility could be achieved by acting on electric loads of appliances, electric loads of heating/cooling systems and energy storage from batteries and electric vehicles. This research aims at developing novel simulation tools for smart cities/smart grid scenarios that exploit the Agent-Based Modelling (ABM) approach to evaluate novel strategies to manage energy in future energy communities. Moreover, the candidate will develop novel Artificial Intelligence algorithms for evaluating the demand flexibility of the consumers by considering users preference and comfort.
Rsearch objectives and methods: The candidate will develop an Agent-Based Platform that will provide a realistic testbed where different management algorithms can be evaluated and compared.
The platform should be based on real data and demand profiles, and should be flexible and extendable so that i) It can be improved with new data from the field; ii) it can be interfaced with other simulation layers (i.e. physical grid simulators, communication simulators); iii) It can interact with external tools executing real policies (such as energy aggregation). The developed platform will help in understanding the impact of new actors and companies (e.g., energy aggregators) in both marketplace and society.

This Agent-Based simulator will be designed and developed to span different spatial-temporal resolutions. For example, spatial resolution would range from the single dwelling up to districts and cities. Whilst, time resolution would range from minutes up to years.

Hence, the research will focus on the development of:
1. agents that express the final customer/prosumer beliefs desire and intention and opinions;
2. tools for modelling the flexibility of each prosumer/consumer;
3. the effects of such flexibility in solving grid constrain problems;
4. Artificial intelligence algorithms for estimating consumers preference;
5. Artificial Intelligence algorithms for estimating the demand flexibility.

All the software entities will be coupled with external simulators of grid and energy sources in a plug-and-play fashion becoming part of the existing co-simulation platform provided by the Energy Center Lab. This will enhance the resulting AMB framework also unlocking hardware in the loop features.

The outcomes of this research will be an agent-based modelling tool that can be exploited for:
- planning the evolution of the future smart multi-energy system;
- evaluating the effect of different policies and related customer satisfaction;
- evaluating the diffusion and acceptance of demand flexibility strategies;
- evaluating the new business models.
Outline of work plan: 1st year. The candidate will study the state-of-the-art solution of existing agent-based modelling tools to identify the best available solution for large scale smart energy system simulation in distributed environments. Furthermore, the candidate will review the state of the art in demand flexibility modelling to identify the challenges and identify possible innovations. Moreover, the candidate will focus on the review of Artificial Intelligence algorithms for estimating demand flexibility and users' preferences. Finally, it will start the design of the overall ABM framework and algorithms starting with the requirements identification and definition.

2nd year. The candidate will terminate the design phase and will start the implementation of the agents' intelligence. Furthermore, it will start integrating the resulting ABM framework with the co-simulation platform provided by the Energy Center Lab to build and test the first version of the proposed solution.

3rd year. The candidate will ultimate the overall ABM framework and AI algorithms and test it in different case studies and scenarios to assess the impact of new energy management strategies and novel business models on the smart grid, the different corporate structures, billing and sharing mechanism in energy communities.
Expected target publications: IEEE Transaction Smart Grid
IEEE Transactions on Evolutionary Computation
IEEE Transactions on Control of Network Systems
Environmental Modelling and Software
Engineering Application of Artificial Intelligence
Journal of Artificial Societies and Social Simulation
ACM e-Energy
Current funded projects of the proposer related to the proposal:
Possibly involved industries/companies:

Title: Co-simulation platform for real-time analysis of smart energy communities
Proposer: Enrico Macii, Lorenzo Bottaccioli, Edoaro Patti
Group website: http://www.eda.polito.it
Summary of the proposal: The emerging concept of smart energy societies and cities is strictly connected to heterogeneous and interlinked aspects, from energy systems to cyber-infrastructures and active prosumers. One of the key objectives of the Energy Center Lab (EC-L), an interdepartmental centre of PoliTo, is to develop instruments for planning current and future energy systems, accounting for the complexity of the various interplaying layers (physical devices for energy generation and distribution, communication infrastructures, ICT tools, market and economics, social). The EC-L tackles this issue by aiming at building a virtual model made of interactive, interoperable blocks. These blocks must be designed and developed in the form of multi-layer distributed infrastructure. Examples of systems realizing partial aspects of this infrastructure have been recently developed in the context of European research projects, such as energy management of district heating systems, smart-grid simulation, thermal building simulation systems, renewable energy source planning. However, a comprehensive and flexible solution for planning and simulating future smart energy cities and societies is still missing. The research program aims at developing the backbone multi-model and multi-energy co-simulation infrastructure allowing to interface and interoperate real/virtual models of energy production systems, energy networks (e.g. electricity, heat, gas), communication network and prosumer models.
Rsearch objectives and methods: This research aims at developing a novel distributed infrastructure to model and co-simulate different Multi-Energy-Systems and general-purpose scenarios by combining different technologies (both Hardware and Software) in a plug-and-play fashion and analyzing heterogeneous information, often in real-time. The resulting infrastructure will integrate into a distributed environment heterogeneous i) data-sources, ii) cyber-physical-systems, i.e. Internet-of-Things devices, to retrieve/send information in real-time, iii) models of energy systems, iv) real-time simulators, v) third-party services to retrieve information in real-time data, such as meteorological information. This infrastructure will follow the modern software design patterns (e.g. microservice) and every single component will adopt the novel communication paradigms, such as publish/subscribe. This will ease the integration of “modules” and the link between them to create holistic simulation scenarios. The infrastructure will enable also both Hardware-in-the-Loop (HIL) and Software-in-the-Loop again to perform real-time simulations. In a nutshell, the co-simulation platform will offer simulations as a service that can be used by different stakeholders to build and analyze new energy scenarios for short- and long-term planning activities and for testing and managing the operational status of smart energy systems. The starting point of this activity will be the already existing EC-L co-simulation platform.

Hence the research will focus on the development of a distributed co-simulation platform capable of:
- Interconnecting and synchronizing digital real-time simulators, even located remotely
- Integrating Hardware in the loop in the co-simulation process
- Easing the integration of simulation modules by automatizing the code generation to build new simulation scenarios in a plug-and-play fashion

The outcomes of this research will be a distributed co-simulation platform that can be exploited for:

- Planning the evolution of the future smart multi-energy system by taking in to account the operational phase
- Evaluating the effect of different policies and related customer satisfaction
- Evaluating the performances of hardware components in a realistic test-bench
Outline of work plan: 1st year. The candidate will study the state-of-the-art solution of existing co-simulation platforms to identify the best available solution for large scale smart energy system simulation in distributed environments. Furthermore, the candidate will review the state of the art in hardware in the loop integration and automatic composability of the scenario code by identifying challenges and possible solution. Finally, it will start the design of the overall platform starting from the requirements identification and definition.

2nd year. The candidate will end the design phase and will start the implementation of the co-simulation platform including HIL features to be integrated with software simulators together to create the first beta version of the tool. Furthermore, the candidate will start developing software solutions to solutions to ease the integration of simulation modules by automatizing the code generation. Moreover, the candidate will submit his/her research to at least an international journal.

3rd year. The candidate will ultimate the overall platform and test it in different case study and scenarios to show all capabilities of the platform in terms of automatic scenario composition and integration of HIL. Moreover, the candidate will submit the follow up of his/her research to at least an international journal.
Expected target publications: IEEE Transaction Smart Grid
IEEE Transaction on Industrial Informatics
Parallel and distributed computing
Environmental Modelling and Software
ACM e-Energy
Current funded projects of the proposer related to the proposal:
Possibly involved industries/companies:

Title: Cybersecurity Automation for Cyber-Physical Systems
Proposer: Riccardo Sisto
Group website: http://netgroup.polito.it
Summary of the proposal: Thanks to the advances of the IoT, cyber-physical systems are now proliferating, and many of them are safety critical. Smart Grids, autonomous driving vehicles, and Industry 4.0 are just some examples of this trend. Being these systems safety-critical, but at the same time open and distributed, cybersecurity is of primary concern for them. At the same time, cybersecurity management is challenging for cyber-physical systems, for several reasons. They are often highly dynamic and reconfigurable (e.g., automotive systems can be configured with many different on-board optional features), and they may have heterogeneous components, including some with restricted computational and storage capabilities.
Taking all these features into account while providing high security assurance is a great challenge the research in this field is trying to address. The complexity of this task makes manual solutions unfeasible and too error-prone, demanding for automation. The candidate will study the problem and develop new cybersecurity automation solutions for cyber-physical systems, capable of adapting to system changes while preserving formally provable security properties. To achieve this goal, the candidate will exploit the latest advances recently made in network security automation and formal methods.
Rsearch objectives and methods: The research work will be especially oriented to identifying innovative lightweight formal modeling techniques that can capture the essential aspects of communication and security configuration in a cyber-physical system, without leading to excessive complexity and intractability.
Modeling may target a variety of aspects such as the communication and cryptographic protocols used in the system, trust relationships, storage and computation capabilities available to run cryptographic operations, and the security-relevant physical properties and states of the systems.
The candidate will pursue correctness-by-construction approaches, alternative to formal verification, but based on the same kind of formal models used for formal verification. Instead of just checking that the behavior of a fully defined model satisfies some properties (verification), some aspects of the model (e.g. the configurations or locations of some security mechanisms) are left open, and a solution that assigns these open values is searched (formal correctness by construction). In order to achieve this goal, Satisfiability Modulo Theories (SMT) solvers will be considered. These tools are good candidates for this purpose because they are very efficient in determining if a set of logical formulas is satisfiable or not, and if it is, they can also efficiently find assignments of free variables of the model that make the logical formulas true. The main idea to be explored in the research is to exploit an SMT solver as backend for building a tool that can automatically and quickly synthesize correct-by-construction security configurations of cyber-physical systems. This implies finding ways of encoding the correct construction problem of such systems as an SMT problem that can be solved efficiently enough. The candidate will exploit the experience already gained by the Netgroup with SMT solvers for fast formal verification and automatic configuration of NFV-based networks.
Outline of work plan: Phase 1 (1st year): the candidate will analyze the state-of-the-art of formal approaches for the modeling and verification of security in cyber-physical systems, and IoT-based network infrastructures, with special attention to the approaches already developed within the NetGroup. Also, recent approaches that go in the direction of correctness by construction with formal guarantees will be studied.
Subsequently, with the guidance of the tutor, the candidate will start the identification and definition of the new approaches for automatic security enforcement based on correctness by construction, as explained in the previous section. At the end of this phase, the publication of some preliminary results is expected. During the first year, the candidate will also acquire the background necessary for the research. This will be done by attending courses and by personal study.

Phase 2 (2nd year): the candidate will consolidate the proposed approaches, will fully implement them, and will make experiments with them, e.g. in order to study their scalability and generality. The approaches will be validated in the domain of industrial networks or automotive systems, by means of specific use cases. The results of this consolidated work will also be submitted for publication, aiming at least at a journal publication.

Phase 3 (3rd year): based on the results achieved in the previous phase, the proposed approach will be further refined and improved, in order to improve aspects such as scalability, performance, and generality of the approach, and the related dissemination activity will be completed.
Expected target publications: The contributions produced by the proposed research can be published in conferences and journals belonging to the areas of security (e.g. IEEE SandP, ACM CCS, NDSS, ESORICS, IFIP SEC, DSN, ACM Transactions on Information and System Security, or IEEE Transactions on Secure and Dependable Computing), cyber-physical systems and related applications (e.g. IEEE Transactions on Industrial Informatics or IEEE Transactions on Vehicular Technology), and networking (e.g. INFOCOM, ACM/IEEE Transactions on Networking, IEEE Transactions on Network and service Management, Netsoft).
Current funded projects of the proposer related to the proposal: ASTRID (AddreSing ThReats for virtualIseD services) H2020 project
Possibly involved industries/companies:In the automotive field, our research group collaborates with Italdesign.

Title: Reliability and Safety of Neural Networks running on GPUs
Proposer: Matteo Sonza Reorda
Group website: http://www.cad.polito.it
Summary of the proposal: Neural Networks (NNs) are increasingly used in many application domains, including some where safety is crucial (e.g., automotive and robotics). In most cases, the NN is executed on a GPU architecture, which allows facing the huge computational power when the NN is deep and complex. Possible faults affecting the hardware of the GPU executing the NN can severely impact the produced results. Unfortunately, we still miss a full understanding of which are the most critical faults, and which are the most sensible modules in a GPU. The goal of the proposed research activity is to fill this gap, leveraging on the available expertise and infrastructures developed by the CAD Group, e.g., in terms of GPU models and Fault Injection environments. A major issue to be faced is how to tame the computational effort required to simulate the effects of the huge number of faults that may affect the hardware, tracing their effects up to the application level. The research activity will focus first on reliability analysis to identify the most critical faults/modules, and then on the development of suitable solutions to increase the reliability and safety, acting both on the hardware and software.
Rsearch objectives and methods: GPUs are increasingly adopted in safety-critical applications (e.g., in the automotive and robotics domains), where the probability of failures must be lower than well-defined (and extremely low) thresholds. This goal is particularly challenging, since GPUs are extremely advanced devices, built with the highly sophisticated (and hence less mature) semiconductor technologies. On the other side, since these applications are often based on Artificial Intelligence (AI) algorithms, they benefit of their intrinsic robustness, at least with respect to some faults. Unfortunately, given the complexity of these algorithms and of the underlying architectures, an extensive analysis to understand which faults/modules are particularly critical is still missing. The planned research activities aim first at exploring the effects of faults affecting the hardware of the GPU implementing the NN. Experiments will study the effects of the considered faults on the results produced by the NN. This study will mainly be performed resorting to fault injection experiments. In order to keep the computational effort reasonable, different solutions will be considered, combining simulation- and emulation-based fault injection with multi-level one. The trade-off between the accuracy of the results and the required computational effort will also be evaluated.
Based on the results coming from the first phase, the second phase of the research activity will aim at devising solutions able to harden the NN execution with respect to hardware faults, acting on the hardware, and/on the software. The impact of the proposed hardening solutions in terms of area and performance overhead will be evaluated, as well as with respect to the number of faults they can detect and/or tolerate.
The challenging objective of this proposal is to identify suitable techniques able to reduce to an acceptable value the probability that faults in a GPU executing a NN produce a critically wrong result in a safety-critical application.
Outline of work plan: The proposed plan of activities is organized in the following phases:
- phase 1: the student will first study the state of the art and the literature in the area of NNs, their implementation on GPUs and their applications. Suitable cases of study will also be identified, whose reliability and safety could be analyzed with respect to faults affecting the underlying hardware.
- phase 2: suitable solutions to analyze the impact of faults on GPUs will be devised and prototypical environments implementing them will be put in place.
- phase 3: based on the results of a set of fault injection campaigns performed to assess the reliability and safety of the selected cases of study a detailed analysis leading to the identification of the most critical faults/components will be carried out.
- phase 4: based on the results of the previous phases, new techniques for NN hardening will be devised and evaluated.
Phases 2, 3 and 4 will also include dissemination activities, based on writing papers and presenting them at conferences. We also plan for a strong cooperation with the researchers of the Federal University of Rio Grande do Sul (Brazil), having special expertise in reliability evaluation, and with NVIDIA engineers.
This proposal is complementary and strictly related to the one titled "Autonomous systems' reliability and safety": both focus on the reliability of AI architectures for safety-critical embedded applications, but the two proposals deal with different underlying hardware architectures.
Expected target publications: Papers at the main international conferences related to test, reliability and safety (e.g., ETS, ATS, VTS, IOLTS, ITC, DSN).
Papers on the main IEEE journals related to design, test and reliability (e.g., IEEE Design and Test, IEEE Transactions on VLSI, IEEE Transactions on CAD, IEEE Transactions on Reliability, IEEE Transactions on Computers)
Current funded projects of the proposer related to the proposal:
Possibly involved industries/companies:NVIDIA

Title: Zero-trust security of network nodes
Proposer: Antonio Lioy
Group website: https://security.polito.it/
Summary of the proposal: Modern ICT infrastructures are characterized by a dissolution of traditional boundaries. Computing and storage are no more available only at the core but – with edge and fog computing, personal devices, and IoT – there are several distributed components that concur to data processing and storage. In a similar way, networks are no more just switching packets through hardware appliances but have evolved into intelligent elements able to perform several tasks. SDN (Software Defined Networking) permits intelligent packet processing based on an external supervisor (the controller and the SDN applications), while NFV (Network Function Virtualization) permits to implement on-demand processing (firewall, VPN, …) once requiring dedicated appliances.
This scenario encourages the adoption of a new security paradigm: zero-trust security. It assumes that no component can be trusted a-priori but it must demonstrate its identity and integrity before being accepted as a member of the infrastructure, as well as periodically during its operation. The research will deal with various aspects of zero-trust security, from electronic identity (of users and devices) to verification of component integrity and monitoring that the intended security properties are built and maintained during system operation. The final objective is the design and test of a coherent zero-trust architecture for modern ICT infrastructures.
Rsearch objectives and methods: The proposed research aims to go beyond the state of the art with respect to the base technologies that support the zero-trust security paradigm.
First, each node has to prove its trustworthiness. This is often achieved by providing an unforgeable proof of its software status (binaries, memory, and configuration). While the basics are covered by the Trusted Computing paradigm (through the remote attestation procedure) there are several aspects to be investigated such as run-time attestation, deep attestation, attestation in virtualized environments, fast attestation, root-of-trust hooks for low-cost devices.
Second, in order to avoid human errors, the system must be configured as much as possible automatically, and its operations must be continuously compared with the expected behaviour. For the networking part, "intents" are currently the hot topic to express the desired behaviour and perform automatic configuration. We plan to extend networking intents to cover also various security aspects (confidentiality, integrity, availability, and privacy) and to use them not only for configuration but also for monitoring the system behaviour.
Last but not least, strong identity verification of all the nodes is needed, for implementing access control and to store reliable and undeniable evidence of the performed actions (so providing support for security audit and even forensic analysis). The strongest identity verification is provided by asymmetric cryptography and PKI, but this has limits in terms of performance (speed and required hardware capability and trust (the hierarchical model is not always acceptable in various environments). Here we will investigate alternative identity solutions, based on fast key distribution, low-power cryptography, and delegated identity (e.g. based on a proxy, reputation, or agreement, possibly backed by an appropriate type of blockchain).
Outline of work plan: The first year will be devoted to gaining insight into the technologies at the foundation of zero-trust security (ZTS): electronic identity (for users and devices) based on PKI and Blockchain, intent-based networking for configuration, and trusted computing for integrity verification.
The second year will be used to explore the design alternatives, achieving a balance between centralized and distributed approaches (i.e. core, edge, and clients). This will include also considering various alternatives for implementing critical functions, in hardware or software (e.g. TPM chip versus firmware-based solutions to implement a TEE, Trusted Execution Environment), reactive versus proactive protection, closed or open ecosystem (e.g. with respect to PKI and blockchain solutions). Target of the second year is the design of at least one (but possibly more) ZTS architectures for various environments, mainly those of the EU-funded projects related to this proposal.
Finally, the third year will be devoted to implementation and experimental evaluation of the proposed architecture(s). This will likely take place within one or more of the EU-funded projects that provide the reference framework for this activity, as each project has aspects related to ZTS, as application, building blocks, or security architecture:
- SPIRS will design an open-source hardware-software architecture for strong node identity, natively supporting integrity verification, hence it will provide components and a test-bed for ZTS
- PALANTIR will use ZTS to protect the ICT the infrastructure itself as part of a SECaaS (SECurity-as-a-Service) approach
- FISHY will use ZTS to protect the whole supply chain of an IT service, from clients to servers, independent of the underlying infrastructure
- ROOT exploits the trusted computing features of ZTS to protect GNSS-based time distribution in a 5G infrastructure
Expected target publications: IEEE Transactions on Networking, IEEE Transactions on Secure and Dependable Computing, IEEE Security, Springer Security and Privacy, Elsevier Computers and Security, Future Generation Computer Systems, ...
Current funded projects of the proposer related to the proposal: Various H2020 projects, mainly SPIRS (to start on October 2021) but also:
- PALANTIR (https://www.palantir-project.eu)
- FISHY (https://fishy-project.eu)
- ROOT (https://www.gnss-root.eu)
Possibly involved industries/companies:Most relevant companies directly involved in this research are HP and Telefonica, but others have interest in different aspects (Atos, Altran, NEC, …)

Title: Cybersecurity in the Quantum Era
Proposer: Antonio Lioy
Group website: https://security.polito.it
Summary of the proposal: Computer engineering research interest in Quantum Computing and Quantum Communication has significantly grown in recent years. Quantum computing will enhance classical computation in many areas such a Finance, AI, Chemistry, and Physics. The impact of this new paradigm on cybersecurity will be significant: the advent of Quantum Computing will endanger current public-key cryptography. Quantum communication provides approaches to enable quantum and classical computers communication, leveraging quantum phenomena such as entanglement. One of the ultimate goals of this field is to build a Quantum Internet along with the current classical one to implement protocols otherwise impossible in a classical scenario. In the short to medium term, Quantum Communication and, in particular, Quantum Cryptography offers technologies that could overcome the security issues coming with the quantum advent. One example is the Quantum Key Distribution (QKD).
The main goal of this research activity is to analyse the techniques coming from both Quantum Cryptography and Post-Quantum Cryptography and apply them to a software infrastructure scenario. This activity involves thorough studies of the possible attacks against quantum technologies and aims at identifying effective mitigations.
Rsearch objectives and methods: Exploring and understanding the main principles behind both Quantum Computing and Quantum Communication is the starting point for the research activity. From a security perspective, knowing the main threats of Quantum Computing, such as Shor's algorithm and its implementation to break public-key cryptographic schemes, is essential.

The main goals of the research activity are the following ones.
1. Analysing Post-quantum (PQC) and Quantum Cryptography (QC) algorithms and protocols. As a result of the quantum advent, two main strategies have been proposed to mitigate the related threats: Post-quantum and Quantum Cryptography. The analysis of both approaches is required to enhance current security solutions with the most suitable strategy for specific domains.
2. Simulating Quantum Cryptography protocols. Special-purpose quantum devices are expensive and require access to extensive network infrastructures for testing. An approach to boost design and test of quantum algorithms and protocols is given by the simulation of the quantum aspects. Several frameworks are available for this purpose (e.g. SimulaQron, NetSquid, Qiskit) and could be used to test the protocols in the scope of Quantum Cryptography.
3. Integrating PQC and Quantum Cryptography with common state-of-the-art security protocols (e.g. TLS, IPsec, SSH) and modern software infrastructures in diverse domains such as Cloud-, Fog-, and Edge-computing, as well as in Network Functions Virtualisation (NFV).
4. Analysing the attacks against PQC and Quantum Cryptography, and the possible countermeasures. PQC suffers from a plethora of side-channel attacks (e.g., timing, fault, cold-boot attacks). Quantum Cryptography and, in particular, QKD leaves room for different kinds of attacks: individual, collective, and coherent (e.g., intercept-resend). In addition, the non-idealities introduced by the implementation of quantum devices lead to specific "quantum attacks" (e.g. PNS, Time-shift attack). Because of this, classical techniques could be used to enhance those systems security (e.g. Privacy Amplification).
Outline of work plan: The first year will be spent studying Quantum Computing, Quantum Communication, and PQC principles, algorithms, and technologies. The PhD student will also analyse modern security paradigms applied to software infrastructures. During this year, the student should also follow most of the mandatory courses for the PhD and publish at least one conference paper. This paper should reflect a preliminary analysis on the application of quantum technologies for the security of software infrastructures.
During the second year, the PhD student will analyse a domain-specific application of Quantum Cryptography, PQC or even a hybrid approach. The application domain should be oriented to modern infrastructures that heavily rely on virtualisation technologies. This analysis will lead to the design of a specific security solution that involves at least one of the aforementioned paradigms. At the end of the second year, the student should have started preparing a journal publication on the topic and submit at least another conference paper.
Finally, the third year will be devoted to the evaluation of the proposed solution. This could be achieved in the case of Quantum Cryptography leveraging simulation platforms, and if possible, actual use case scenarios utilising quantum devices. This evaluation and experimental phase could be complemented by cooperation with other departments inside POLITO, also leveraging some ongoing projects in which an experimental facility and tests on physical devices are expected. A promising project for this purpose is a collaboration with TIM, which focuses on exploring Quantum Communication technologies and experimenting with QKD protocols both at simulation level and with physical devices. At the end of this final year, a publication in a high-impact journal shall be achieved.
Expected target publications: IEEE Security, Springer Security and Privacy, Elsevier Computers and Security, Future Generation Computer Systems, IEEE Transactions on Quantum Engineering
Current funded projects of the proposer related to the proposal: A project in collaboration with TIM focusing on Quantum Communication and, in particular, on QKD simulation.
Possibly involved industries/companies:Interest in this activity from TIM and Telefonica, although there is not yet any direct formal involvement.

Title: Evaluating the impact of automated decision systems in urban contexts
Proposer: Juan Carlos De Martin, Francesca Governa, Antonio Vetrò
Group website: https://nexa.polito.it/
Summary of the proposal: Nowadays automated decision (and recommendation) systems are used in several domains such as penal justice, advertising, credit scoring. They take or suggest significantly impacting decisions on the life of people, and it has been widely reported that these tools can cause discriminations against certain population groups (e.g., people with dark skin, females, etc), due to bias incorporated in the datasets used by these tools, or in the algorithms themselves [1]. While law is still failing in regulating the usage of automated decision systems, journalists and researchers collected relevant evidence about the problem, however consensus among mitigation strategies is still far from being reached.
Among the cases reported in the literature, many concern applications of such systems in urban contexts, where physical spaces and characteristics of population influence each other [2]. The PhD candidate will investigate automated decision systems in urban context to understand whether they amplify or reduce existing inequalities in cities, or they create new ones.

[1] O’Neil, C. (2017). Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy . New York: Broadway Books.

[2] Wacquant L. (2016), Revisiting territories of relegation: Class, ethnicity and state in the making of advanced marginality, Urban Studies, 53(6) 1077–1088.
Rsearch objectives and methods: The PhD proposal is in collaboration with the Future Urban Legacy Lab. The general goal of the PhD proposal is to understand whether automated decision systems in urban context amplify or reduce existing inequalities in cities, or they create new ones.

The objectives are the following ones:

O1: Mapping current demographic disparities
A preliminary step in the research is to study inequalities in the urban field, understanding influencing factors, how it is measured, how it is represented. These are aspects of interdisciplinary research that will require strict collaboration with the Future Urban Legacy Lab.

O2: Understanding the relationship between inequalities in cities and the usage of automated decision systems
This objective should be reached with a two-fold approach:
i) an analysis of scientific evidence and journalistic investigations to understand the current documented impact of automated decision systems in urban context (e.g., digital welfare, predictive policing, risk scores)
ii) experimentation/simulation with real/synthetic dataset

O3: Design, implement and test alternative tools
The PhD candidate should design, implement and test remediation strategies to the main issues collected in previous step. The main expected outcomes are alternative algorithms or data curation processes techniques that lessen the disparate impact of classification and/or prediction tasks. In addition, the candidate shall elaborate qualitative insights and policy suggestions on how to shape the digital infrastructure and the processes for decision making software, and to elaborate scenarios on different domains.
Outline of work plan: O1: Understanding inequalities in cities with an interdisciplinary approach
- A1.1 Elaboration of a conceptual and operational data measurement framework to measure inequality in the urban field
- A1.2: Definition of a set of guidelines for visualizing inequalities in the urban field
- A1.3: Implementation of a few examples of urban inequality maps following guidelines

O2: Understanding the relationship between inequalities in cities and the usage of automated decision systems
- A2.1 Literature review on the impact of automated decision systems in urban context
- A2.2 Elaboration of a conceptual and operational data measurement framework for identifying data input characteristics that potentially affect the risks of discriminating software decisions.
- A2.3 Experimentation with (or simulation) of an existing (or hypothetical) automated decision system, which implies classification or prediction tasks on specific population groups, and analysis of impact, also in relation to different fairness metrics already established in the scientific literature

O3: Mitigation strategies
- A3.1 Based on previous activities findings, design and implementation of mitigation and remediation strategies to reduce the negative impact of classification and/or prediction tasks.
- A3.2 Complement with explanations and critical reflections in the context of digital infrastructure for the future of cities, in relation to ongoing activities at the Future Urban Legacy Lab.
Expected target publications: Illustrative examples of targeted scholarly journals include:

Transactions on Software Engineering and Methodology
ACM Transactions on Information Systems
Expert Systems with Applications
Government Information Quarterly
Journal of Systems and Software
Software Quality Journal
Big Data and Society
Big Data Research
Significance
Daedalus
Information sciences
IEEE Access
IEEE Computer
IEEE Software
Communications ACM
Current funded projects of the proposer related to the proposal:
Possibly involved industries/companies:Potential involvement of the following industries/organizations: TIM, ISO/IEC