Corsi Anno Accademico 2022/2023


Social Media Manipulation

The most important outcomes for the audience will be:

1. A general understanding of the main research problems and most relevant findings in the new ‘science of fake news’ multidisciplinary sub-field, focusing on a selection and unavoidably biased list of some of the most influential and related scientific papers that have been published so far, especially in the last 8 years;
2. An introduction of the underlying mechanisms that make fake news propagation fast and difficult to stop (such as homophily, segregation and polarization in social networks, belief reinforcements, simple vs. complex social contagion, and so on);
3. The main methodologies used for understanding the phenomenon, as well as some of the techniques adopted so far to try to limit the spreading of low-quality information (and why they basically fail in the short term). These methodologies are based on deterministic as well as not deterministic modeling that allows what-if analyses as well as on data-driven approaches based on empirical observations;

Although the set of problems and methodologies is naturally multidisciplinary, the main field
category is computer science; nevertheless, anyone willing to make data science their own
field of expertise would be potentially interested in this topic.

Emotion Oriented Systems

Aggregate Programming for the Internet of Things (aa 2023-2024)

Knowledge management and information extraction from structured and unstructured data for process mining (techniques, algorithms, and tools)

The discipline of Process Mining has recently captured increasing attention, extending from organizational and management studies to all areas where data with timed events are present.

The automated analysis of timed events benefits from a broad set of methods, techniques and tools capable of extracting information from structured and unstructured data.

The PhD course will provide an overview of main topics in the Process Mining discipline to discover and analyze temporal processes, including practical techniques and tools for visualizing processes, identifying bottlenecks, performing variant analysis, introducing predictive process monitoring, comparing time series data in a ‘conformance checking’ perspective, processing data to extract information from text in event log format.

Applications are considerable to "processes" of very different types, e.g., educational, healthcare, legal, chatbot processes, and so on.

Scheduling of course lectures:

  • I lesson - May 17 - 2:30 p.m. - 5:30 p.m. (Third floor, Sala riunioni)
    LINK Webex: 

  • II lesson - May 21 - 2:30 p.m. - 5:30 p.m. (Third floor, Sala riunioni)
    LINK:

  • III lesson - May 23 - 2:30 p.m. - 5:30 p.m. (Third floor, Sala riunioni)
    LINK:

  • IV lesson - Jun 5 - 10 - 13 (First floor, Sala Seminari)
    LINK:

  • V lesson - Jun 6 - 10 - 13 (First floor, Sala Seminari)
    LINK:

  • VI lesson - TBA

  • VII lesson - TBA
  • VIII lesson - Jun 6 - 14-17 p.m. (First floor, Sala seminari)

Program

The PhD course consists of seminars alternating theory and practice carried out by the following teachers:

Emilio Sulis - Introduction to Process Mining algorithms, techniques, and tools (pm4py/ProM/bupaR/DISCO)
Luigi Di Caro - Knowledge extraction from textual data, event log enrichment
Laura Genga (Technical University of Eindhoven) - Variant Analysis, Conformance checking techniques
Chiara Difrancescomarino (University of Trento) - A discussion on AI and PM

The course introduces the main techniques for process discovery, validation and improvement from event-logs, typically extracted from information systems, sensors, web applications.

"Predictive process monitoring" and NLP techniques for feature set extraction and enrichment will be introduced, including a combination of data mining, text mining, and process analysis.

Teaching material

Course material will be provided by the teachers. There are no required textbooks for this course.
The lecturers will propose papers, documentation and websites as educational materials during the course.

Learning assessment methods

The examination consists of a written presentation (e.g., a set of slide or a short paper) concerning the student's preferred argument/topic.
It is also possible to analyse data of interest, as well as use case studies/dataset provided by the lecturers.

Suggested readings and bibliography

(suggested) Wil M. P. van der Aalst, Josep Carmona: Process Mining Handbook, Springer 2022, ISBN 978-3-031-08847-6 - https://link.springer.com/book/10.1007/978-3-031-08848-3

Methods and tools for HPC and convergent computing

Type theory at work: introduction to Agda

The course aims to introduce formal computer science methods using tools based on type theory, especially AGDA. Lectures will concern the grounds of type theory, namely intuitionistic logic and lambda calculus, eventually focussing on dependent types and Martin-Loef intuitionistic type theory, of which AGDA is an implementation.

Cloud Computing for Science 2024

See https://dott-informatica.campusnet.unito.it/do/corsi.pl/Show?_id=gu2e

Scheduling in Linux and Real-Time Systems

The course is composed by a standard part and a customized part.

Standard part:

  • model of tasks, machines, and schedulers
  • real-time constraints
  • popular scheduling algorithms and tests for real-time tasks on a single processor
  • popular scheduling algorithms and tests for real-time tasks on a multi processor
  • virtual processors

The customized part will be decided in agreement with the participating students. It may include selected topics among:

  • Linux
    • the scheduler, setting the desired scheduler
    • tracing events
  • KVM
    • virtualization with KVM
  • Communication along task chains
  • basics of Zephyr RTOS

Introduction to Deep Learning 2024