| 1 | Scalable data processing for data science (16 hrs, 4 ECTS) |
| LECTURER: Paolo MISSIER, University of Birmingham, UK SYLLABUS: The teaching offer is designed to fit within the context of Unimore’s PhD programs in Computer Science, Computer Engineering, and Math. It will consist of a series of topical seminars and practical lab sessions, for a total of about 16 hours. The seminars are organized into two parts. Part I, described here, runs over 10 hours and is focused on Explainable AI (XAI), covering the following topics: – The need for explanations in Machine Learning and AI – Classical global and local methods : LIME and Shapley (Shap values) – Influence functions: a robust statistical approach that has recently (circqa 2017) been revisited to associate a relative importance to training examples for explaining a specific model inference. One key advantage of using influence analysis over other methods, is its ability to bypass the “black box” barrier that is typical of complex nonlinear models (eg deep neural networks). The programme includes: – About 4 hours of seminar-style lectures, using key papers and a recent survey to dive into specific approaches to using influence functions – Guided study to relevant literature, where students are required to select recent work from a portfolio of recommended papers, and to present their insights to the class – Practical work using python libraries that implement influence-based methods to provide explanations, in combination with simple ML models. DATES: TBD | |
| 4 | Privacy-enhancing technologies (12 hrs, 3 ECTS) |
| LECTURER: Luca FERRETTI, UNIMORE – FIM SYLLABUS: The course offers an introduction to privacy-enhancing technologies (PET), which are security solutions that aim at minimizing information disclosure during data processing. The course outlines the most important system models and security guarantees related to PETs, including information sharing, collaborative and outsourced computation, transparency architectures, and discusses the most popular techniques based on applied cryptography and hardware-related technologies. The course especially focuses on practical solutions and include hands-on sessions based on existing software frameworks. DATES: TBD | |
| 5 | Advanced GPU programming – libraries for data science (16 hrs, 4 ECTS) |
| LECTURER: Nicola CAPODIECI, Filippo MUZZINI, Roberto CAVICCHIOLI, UNIMORE – FIM/DCE SYLLABUS: This course will dive deep into advanced concepts of GPU Programming and architectures. In the first part, advanced CUDA programming concepts will be presented, focussing on the latest evolution of the CUDA programming model and architectural features such as scheduling, programming best practices and CPU – GPU interaction optimizations along with comparison discussions with other GPGPU APIs. In the second part of the course, widely used CUDA libraries for numerical calculus and data science will be presented through extensive examples, such as cuBLAS, cuSOLVER and cuSPARSE for numerical optimization, cuFFT for signal processing and NPP for image processing. DATES:TBD | |
| 6 | Internet and Web of Things at the Edge (8hrs, 2 ECTS) |
| LECTURER: Luca BEDOGNI, UNIMORE – FIM SYLLABUS: The class will focus on modern smart IoT systems, which are found in many different scenarios. The course will present at first the Internet of Things and Web of Things scenario, and will then move to practical problems in such environments. The class is focused on the key challenges that low power devices have when processing information. This can be done directly on the device itself, if possible, or can be offloaded to close edge servers. Two reference scenarios pertaining e-Health and Industry 4.0 will be discussed and analyzed in detail, highlighting the key differences and practical issues which can be found. The class will conclude by discussing future research directions. DATES: TBD | |
| 7 | Vulnerability research (12 hrs, 3 ECTS) |
| LECTURER: Mauro ANDREOLINI, UNIMORE – FIM SYLLABUS: This course will introduce students to the methodological and practical aspects of security vulnerability research in software. The involved activities are at the basis of identifying unknown and sophisticated flaws (typically resulting in a chain of more elementary ones) that current software solutions are not able to find automatically. Classes are organized as a series of seminars and practical lab sessions discussing the following topics: – static analysis – dynamic analysis – formal verification methods – semi-automated testing DATES: TBD | |
| 8 | Efficient DL/ML models for embedded systems (12hrs, 3 ECTS) |
| LECTURER: Alessandro CAPOTONDI, UNIMORE – FIM SYLLABUS: The execution of sophisticated Artificial Intelligence (AI) workloads is no longer a prerogative of high-end, high-performance computing systems. Energy- and resource-constrained embedded devices, also called edge devices, are increasingly embracing this type of functionality, which is key to enabling the realization of smart, autonomous systems (unmanned aerial and terrestrial vehicles, robotic arms, etc.). This class will present an overview of the state-of-the-art methodologies for effective and efficient deployment of Deep Learning and Machine Learning (DL/ML) models on edge embedded systems. The topics presented include quantization, pruning and network-architecture-search (NAS) strategies targeting practical and realistic challenges of deploying state-of-the-art DL/ML tasks on edge systems (e.g. NVIDIA Tegra, Xilinx MPSoC, MCU-class RISC-V and ARM SoCs). The class will then close by providing insights on near-to-come research directions in the field. DATES: TBD | |
| 9 | Complexity Theory, On-Line and Approximation Algorithms (12 hrs, 3 ECTS) |
| LECTURER: Manuela MONTANGERO, Mauro LEONCINI, UNIMORE – FIM SYLLABUS: The course will introduce students to theory of computational complexity. The first (shorter) part of the course will be dedicated to the introduction of the fundaments of complexity theory and the definition of the most important complexity classes. The second part of the course will be dedicated to the study of approximation and on-line algorithms. The former are used to address difficult problems (NP-complete or NP-hard), the latter for those problems whose input is not completely available at the beginning of the execution of a solving algorithm. To this aim, problems with interesting applications in distributed/parallel system and data science scenarios will be selected. DATES: TBD | |
| 10 | Introduction to Formal Verification: From Foundations to AI and Multi-Agent Systems (16 hrs, 4 ECTS) |
| LECTURER: Angelo FERRANDO, UNIMORE – FIM SYLLABUS: The course introduces PhD students to the fundamental concepts and methodologies in formal verification, focusing both on static approaches such as Model Checking and dynamic techniques like Runtime Verification. Besides exploring theoretical foundations, the course emphasizes practical knowledge through the use of established verification tools in laboratory sessions. A particular focus will be given to formal verification techniques applied to complex AI-driven and Multi-Agent Systems (MAS). Students will gain practical insights into verifying software components and tackling verification challenges arising in distributed and autonomous systems. At the course’s conclusion, recent research advancements and open challenges in runtime verification and verification of multi-agent systems will be discussed, providing students with a comprehensive understanding of current research directions. DATES: TBD | |
| 11 | Introduction to complex systems (8hrs, 2 ECTS) |
| LECTURER: Marco VILLANI, UNIMORE – FIM DATES: TBD | |
| 13 | Multi-objective Optimization and Symbolic Regression (12hr, 3ECTS) |
| LECTURER: Veronica GUIDETTI, UNIMORE – FIM DATES: TBD | |
| 15 | Methods for social and economic science and data (24 hrs, 6 ECTS) |
| LECTURER: Stefano GHINOI, Elvira PELLE, UNIMORE (DCE) SYLLABUS: The course aims at introducing Social Network Analysis (SNA) to doctoral students, which is based on the use of quantitative tools for mapping and analyzing qualitative models of relationships that connect individuals, organizations and institutions. The course provides an overview of the main networking approaches and is structured around a series of theoretical sessions and practical (mini) workshops; in these labs students will have the opportunity to use Python to analyze real networks. The main topics covered in this course are the following: 1) History of SNA and theoretical approaches; 2) network structure data; 3) network statistics; 4) clusters and online communities; 5) network models. By the end of the course, students will be able to understand how to collect, analyze and interpret network data to address social and economic challenges. DATES: TBD | |
| 16 | Quantitative and formal modeling of historical sciences. An introduction to the Parametric Comparison Method (12 hrs, 3 ECTS) |
| LECTURERS: Cristina GUARDIANO, UNIMORE – DCE — Giuseppe LONGOBARDI, UNIVERSITY OF YORK SYLLABUS: The need to reach progressively more profound levels of chronological depth in the investigation of the human past is a requirement for any discipline with ambitions of historical reconstruction. In contemporary times, the achievements reached by historical sciences (e.g., population genetics) in the search for long-persistence patterns able to reveal deep-time relations were possible thanks to two radical paradigm shifts: the adoption of quantitative modeling and automatic procedures, to process and measure big amounts of data and extract generalizations sustained by statistical support, and a qualitative change in the type of taxonomic data, thanks to the discovery that abstract entities, not directly observable but responsible of several variable surface traits, are more able to retain historical information than observable patterns. In linguistics, the development of the historical paradigm in the XIX century has prompted an extraordinary progress in our knowledge about human history by revealing relations among languages/populations which could have not been discovered by archaeology or demography alone, thanks to the identification of abstract patterns of language transmission and change. In the past 30 years, thanks to the development of Quantitative Phylogenetics, historical investigation in linguistics has benefited from the adoption of computer-based techniques, taxonomic algorithms and methods proper of data science, leading to the implementation of a wide array of automatic tools to generate computer-based taxonomies, explore dynamics of language evolution, reconstruct ancestral states and migration patterns, compare linguistic, genetic, and cultural evolution, model language contact, reconstruct character-by-character the evolution of a family from the assumed shared ancestor. These tools prompt excellent results in performing accurate objective reconstructions but has also reveal important limits in attaining the chronological depth required for long-range investigation, demonstrating that the goal of discovering deep-time relations using languages can only be pursued through the combination of quantitative modelling with a radical qualitative change in the level of linguistic characters employed for taxonomic reconstruction. The Parametric Comparison Method (PCM) implements a comparative model precisely based on these tenets. One of its major goals is working out computable tools for assessing historical relatedness between languages against chance when etymological evidence is missing. To this end, the PCM exploits cognitive parametric theories to measure grammatical diversity and its distribution, and demonstrates that abstract cognitive entities retain a significant historical signal able to reveal unknown historical crosslinguistic connections. In the parametric framework of cognitive biolinguistics, human grammars are represented as finite strings of binary values/states (1/0, or +/-). In this approach, the label “parameters” refers to a set of open choices between binary values, generated by our invariant universal language faculty, and closed by each language learner based on the linguistic evidence s/he is exposed to. Parameter systems exhibit two layers of deductive structure: (a) each parameter is responsible for a set of different co-varying surface linguistic patterns (manifestations), and (b) parameters form a network of partial implications: one value (though not the other) of a given parameter p1 may entail the irrelevance of another parameter p2, whose manifestations would then become predictable. The parameter setting algorithm presented in this course is based on all such properties, and consists of the following components: (i) a list of binary parameters; (ii) a list of formulas which define cross-parametric implications in this set; (iii) for each parameter, the list of surface manifestations it generates; (iv) a list of YES/NO questions associated to each manifestation, which are used to collect the data required to set the value of each parameter in a given language (only YES answers set the value 1). DATES: TBD | |
| 18 | Sociology of innovation: new technologies and organization (12hrs, 3 ECTS) |
| LECTURER: Matteo RINALDINI, UNIMORE – DCE SYLLABUS: The course aims to reflect on the relationship between technological innovation and organisational innovation through the various perspectives that, in the socio-economic field and in the field of innovation studies, have subjected a deterministic interpretation of technology to criticism. Various theoretical perspectives derived from organisational studies and innovation studies (SCOT, ANT, sociomateriality, etc.) will be analysed and compared with each other, and through these perspectives the current techno-organisational developments attributable to the so-called fourth industrial revolution and in general to the processes of digitalisation of work and production activities will be analysed. The various topics will be addressed not only through lectures, but also through the discussion of materials, seminars and workshops that may include the presence of external experts and colleagues. DATES: TBD | |
| 19 | Sociosemiotic analysis and Sociology of Data. Examples from Data visualization to environmental and social phenomena (12 hrs, 3 ECTS) |
| LECTURER: Federico MONTANARI, UNIMORE – DCE SYLLABUS: The course analyzes the effects that data and technologies have on current social systems. Through studies in the sociology of sciences and technologies, the course provides a point of view that helps to balance technical-computational aspects with issues of social responsibility. The topics of the course concern the use of data from a technological and social point of view. The sociological and sociosemiotic areas of interest are STS (Science and technological studies) and ANT (Theory of actor networks). DATES: TBD | |
| 21 | Corsi Di Formazione Complementare Per Dottorandi E Assegnisti Ediz. 2024/2025 (24 hrs, 6 ECTS) |
| LECTURER: Barbara REBECCHI, Ferdinando DI MAGGIO, Federica MANZOLI, Nadja SEDING, Giulia CATELLANI, Valeria BERGONZINI, Valeria GOLDONI, UNIMORE – International Research Office SYLLABUS: The course is composed of 4 modular sessions: a. Policies for research and innovation: this session explains where the fundings for research come from. Opportunities and practices for national and international funding for research and innovation are explained; b. Planning the research: In this session all the various phases of the planning of research are explained: the EU finding policies and calls; the project cycle, the structure of the action and cost plan, the actors involved; the negotiation and the management of the european projects; c. Exploitation of the research results; d. Intellectual property: IP rights, protection methods, patents; management and exploitation of the IP; patent databases. This course is mandatory for all PhD students DATES: TBD | |
| 22 | Bibliographic research, scientific writing and dissemination: tools, techniques and strategies (12 hrs, 3 ECTS) |
| LECTURER: Michele POLA, Andrea SOLIERI, UNIMORE – Ufficio Bibliometrico – SBA SYLLABUS: The course aims at teaching the skills and the knowledge for using the specific library services and resources for doctoral students; for being productive in information retrieval, in preparing a bibliography, in writing a scientific article, so to support the Ph.D. students in their path with an outlook at their post-doc career. The course will provide an in-depth introduction to the following aspects: • resources for basic information retrieval like OPACs, UNISTORE and OneClick discovery tool • resources for advanced information retrieval like scientific databases like Scopus and Web of science. • the scientific journal as the main vehicle for STEMM research dissemination. • workflow of a scientific article; (copy)rights and dues of an author, plagiarism, citations and bibliographies. • How to work with a reference manager software. • Improving research impact: from bibliometric analysis to research evaluation; journals evaluation (from authoritative to predatory); differences among ahead of print, post print, editorial version of a publication; • the Open Access initiative and what this means for the authors; what ASN and VQR are with examples and exercises; how the repository IRIS works; checking author profiles on Scopus, Publons and ORCID. This course is mandatory for all PhD students DATES: TBD |
