This report concerns the first year of delivery of the National Doctorate in Artificial Intelligence. We will therefore talk about the students of the 37th cycle, who completed their first year.
The initial effort, conducted in synergy with the other Doctorates in Artificial Intelligence federated with ours, consisted in outlining the fundamental principles for the training and research activities of the doctoral students.
- Structuring of teaching activities
Teaching obligations: minimum 140 hours in total. The activities that students of the national PhD course in Artificial Intelligence (PhD-AI.it) are normally required to carry out include:
- primary educational activities: the attendance and passing of the relative exam, of three or more courses whose hours of frontal lessons are at least 80. These courses are preferably chosen among those included in the didactic offer of our doctoral course and in those of the other 4 doctorates of PhD-AI.it
- additional training activities: attendance of at least 60 hours (in addition to those of the primary activities) of training activities specifically provided for the doctoral courses by the university or research institutions of the PhD-AI.it (or by other Italian or foreign institutions , subject to the approval of the college) in the context of additional activities: attendance of two PhD-AI.it doctoral schools, included in the 60 hours mentioned above
The PhD student presents a study plan (by January) interacting with the supervisors (tutors) and with the coordinator, the college approves. It can be modified during construction, always in agreement with the supervisors. The teaching staff can however establish, for the specific student, a greater number of preparatory courses with exam (up to a maximum of 60 additional hours), in order to fill any gaps.
Exam conclusion: preferably within the first year, without fail within the second.
- Cycles of seminars, "summer school" and other training activities
The following contribute to the achievement of 60 hours of additional training activities:
the cycles of seminars and doctoral schools organized by university or research institutions external to PhD-AI.it (subject to authorization by the college), with the specific indication that they are aimed, exclusively or mainly, at doctoral students.
The following also contribute to the achievement of 60 hours of additional training activities, up to a maximum of 20 hours in total:
- teaching activities concerning research management, knowledge of European and international research systems and the valorisation of research and intellectual property organized by the university or research institutions of the PhD-AI.it or even external to the PhD-AI.it (subject to the approval of the college).
- Structuring of research activities
Individual panel for each PhD candidate: Co-supervision to cover multi-disciplinary needs Panel formation (recommendations):
at least 2 supervisors, at least 3 members in addition to the supervisors, overall, at least 3 members must be part of the College consider diversity of location and disciplines approval of the college.
- Evaluation steps of the PhD course:
Thesis proposal (first year) - submission of the proposal to the panel + presentation and evaluation
Thesis progress (second year) - presentation of the results achieved to the panel
Final thesis (third year) - presentation of the thesis to the panel, evaluation by two external referees
Defense and final evaluation for the achievement of the title
In all three cases discussion with the Panel and approval by the College by the end of October of the first/second/third year
The activity of the first year doctoral students, as per the regulation, was mainly oriented towards the acquisition of training credits.
The PhD organized a doctoral school which was held in the week of 12 September at DIAG (Rome), was attended by all students and saw the provision of the following courses:
• Lecturer: Christine Bauer
Title: Recommender systems: Evaluating comprehensively and assessing the impact
• Lecturer: Michele Flammini
Title: Coalition formation games
• Lecturer: Andre Martins
Title: AI and Natural Language Processing
• Lecturer: Luca Viganò
Title: Formalizing the Human Dimension of Cybersecurity
All students were admitted to the second year.
Alongside the training activities, the PhD students have carried out research activities in the area of their interest, typically publishing the results obtained in international forums (conferences, journals or workshops). The details relating to the research area and the panel of each PhD student are specified below:
Dott. ANGIONI DANIELE
Attività di ricerca: Security Issues of Machine Learning in the Real World
Supervisore: Battista Biggio (Univ. Cagliari),
Co-supervisore: Fabio Roli (Univ. Genova)
Membri: Giorgio Fumera (Univ. Cagliari), Luca Oneto (Univ. Genova), Alessio Merlo (Univ. Genova)
Dott. AVELLA FRANCESCO
Attività di ricerca: Neural databases
Supervisore: Fabrizio Silvestri (Univ. Sapienza)
Co-supervisore: Giovanni Trappolini (Univ. Sapienza)
Membri: Chiara Di Francescomarino (FBK), Riccardo Lazzeretti (Sapienza), Iacopo Masi (Univ. Sapienza)
Dott. BACCIU ANDREA
Attività di ricerca: Retrieval Augmented Neural Databases:
Supervisore: Fabrizio Silvestri (Univ. Sapienza)
Co-supervisore: Nicola Tonellotto (Univ. Pisa);
Membri: Maurizio Lenzerini (Univ. Sapienza), Roberto Capobianco (Univ. Sapienza), Daniele Nardi (Univ. Sapienza).
Dott. BARBIN AYMERIC
Attività di ricerca: Adversarial Machine Learning
Supervisore: Alfonso Gerevini (Univ. Brescia)
Co-supervisore: Federico Cerutti (Univ. Brescia)
Membri: Andrea Micheli (FBK), Andrea Loreggia, (Univ. Brescia), Ivan Serina (Univ. Brescia)
Dott. BEJGU ANDREI STEFAN
Attività di ricerca: Lexical Semantics in Natural Language Processing
Supervisore: Roberto Navigli (Univ. Sapienza)
Co-supervisore: Paola Velardi (Univ. Sapienza)
Membri: Giuseppe De Giacomo (Univ. Sapienza), Bernardo Magnini (FBK), Riccardo Lazzeretti (Univ. Sapienza)
Dott. CAPPELLINI GUGLIELMO
Attività di ricerca: AI-based methods for PDEs with application to Medical Robotics Supervisore: Marilena Vendittelli (Univ. Sapienza)
Co-supervisore: Giovanni Trappolini (Univ. Sapienza)
Membri: Daniele Nardi (Univ. Sapienza), Riccardo Lazzeretti (Univ. Sapienza), Pierfrancesco Pavoni (Medlogix)
Dott. CAPOZZI GIANLUCA
Attività di ricerca: Adversarial Attacks against Binary Similarity Systems
Supervisore: Antonio Di Luna (Univ. Sapienza)
Co-supervisore: Irene Amerini (Univ. Sapienza)
Membri: Lazzeretti Riccardo (Univ. Sapienza), Alessio Merlo (Univ. Genova), Fabio Roli (Univ. Genova)
Dott. HUGUET CABOT PERE-LLUIS
Attività di ricerca: Relation and Information Extraction
Supervisore: Roberto Navigli (Univ. Sapienza)
Co-supervisore: Daniele Nardi (Univ. Sapienza)
Membri: Paola Velardi (Univ. Sapienza), Bernardo Magnini (Univ. Sapienza), Massimo Mecella (Univ. Sapienza)
Dott. LANZINO ROMEO
Attività di ricerca: Analysis of emotions through Deep Learning.
Supervisore: Francesco Scarcello (Univ. Calabria),
Co-supervisore: Luigi Cinque (Univ. Sapienza),
Membri: Danilo Avola (Univ. Sapienza), Iacopo Masi (Univ. Sapienza), Fabio Roli (Univ. Genova)
Dott. LEDDA EMANUELE
Attività di ricerca: Trustworthy AI Through Uncertainty Quantification
Supervisore: Fabio Roli (Univ. Genova)
Co-supervisore: Giorgio Fumera (Univ. Cagliari)
Membri: Enrico Giunchiglia (Univ. Genova), Battista Biggio (Univ. Cagliari), Luca Oneto (Univ. Genova)
Dott. LONGO GIACOMO
Attività di ricerca: Evaluating security properties of complex cyber-physical systems
Supervisore: Alessio Merlo (Univ. di Genova),
Co-supervisore Enrico Russo (Univ. di Genova)
Membri: Battista Biggio (Univ. di Cagliari), Enrico Giunchiglia (Univ. di Genova), Francesco Palmieri (Univ. di Salerno)
Dott. MANCINO ALBERTO CARLO MARIA
Attività di ricerca: Differentially Private Machine Learning Models and their Robustness To Adversaries
Supervisore: Tommaso Di Noia (Politecnico di Bari),
Co-supervisore: Eugenio Di Sciascio (Politecnico di Bari),
Membri: Paolo Mori (CNR), Fabio Roli (Univ. di Genova), Francesco Scarcello (Univ. della Calabria)
Dott. MATHEW JERIN GEORGE
Attività di ricerca: Trustworthy Artificial Intelligence
Supervisore: Donatella Firmani (Univ. Sapienza)
Co-supervisore: Massimo Mecella (Univ. Sapienza)
Membri: Tommaso Di Noia (Poli Bari), Maurizio Lenzerini (Univ. Sapienza), Bernardo Magnini (FKB)
Dott. PIRAS GIORGIO
Attività di ricerca: Adversarial Machine Learning
Supervisore: Battista Biggio (Univ. Cagliari)
Co-supervisore: Giorgio Giacinto (Univ. Cagliari)
Membri: Fabio Roli, (Univ Genova), Enrico Giunchiglia (Univ. Genova), Davide Anguita (Univ. Genova)
Dott. QUARTA ALESSANDRO
Attività di ricerca: Artificial Intelligence technologies for the compliance of Electronic Control Panels
Supervisore: Nicola Leone (Univ. Calabria)
Co-supervisore: Francesco Calimeri (Univ. Calabria)
Membri: Diego Calvanese (Univ. Bolzano), Enrico Giunchiglia (Univ. Genova) Viviana Mascardi (Univ. Genova)
Dott. RAGNO ALESSIO
Attività di ricerca: Topology-based Explanations for Neural Networks
Supervisore: Roberto Capobianco (Univ. Sapienza)
Co-supervisore: Daniele Nardi (Univ. Sapienza)
Membri: Fabrizio Silvestri (Univ. Sapienza), Alfonso Gerevini (Univ. Brescia), Francesco Scarcello (Univ. Calabria)
Dott. ROSSETTI NICHOLAS
Attività di ricerca: Deep Learning techniques to tackle the Covid19 pandemic
Supervisore: Alfonso Gerevini (Univ. Brescia)
Co-supervisore: Ivan Serina (Univ. Brescia), Roberto Capobianco (Univ. Sapienza), Federico Cerutti (Univ. Brescia), Luca Putelli (Univ. Brescia)
Dott. SCIRÈ ALESSANDRO
Attività di ricerca: Automatic text summarization
Supervisore: Roberto Navigli (Univ. Sapienza)
Co-supervisore: Paola Velardi (Univ. Sapienza)
Membri: Giuseppe De Giacomo (Univ. Sapienza), Bernardo Magnini (FBK), Riccardo Lazzeretti (Univ. Sapienza)
Dott. SILO LUCIANA
Attività di ricerca: Modelling Digital Twins Composition
Supervisore: Giuseppe De Giacomo (Univ. Sapienza)
Co-supervisore: Massimo Mecella (Univ. Sapienza)
Membri: Diego Calvanese (Univ. Bolzano), Francesco Scarcello (Univ. Calabria), Francesco Leotta (Univ. Sapienza)
Dott. TEDESCHI SIMONE
Attività di ricerca: Combined Neural and Knowledge-based Data Creation for Multilingual Named Entity Recognition
Supervisore: Roberto Navigli (Univ. Sapienza)
Co-supervisore: Daniele Nardi (Univ. Sapienza)
Membri: Paola Velardi (Univ. Sapienza), Bernardo Magnini (Univ. Sapienza), Massimo Mecella (Univ. Sapienza)