Informatica - Attività didattica

Piano didattico – 35°ciclo

Calendario lezioni - Google Calendar


Elenco degli insegnamenti offerti nell’a.a. 2019/2020

From Expert Systems to Wearable Expert Systems: Definitions, Models, and Future Challenges
Teacher  Fabio Sartori – Dept. of Informatics, Systems and Communication – University of Milano-Bicocca
Language English
CFU  2.5
Hours  20
Program  The course will discuss about the definition of Expert Systems over the last four decades in order to understand how this kind of applications have adapted to the technological progress in IoT research field.

Indeed, the recent availability of more and more sophisticated wearable devices at relatively low costs opens new frontiers in the expert systems research field. Applications that can be referred to as «expert systems» run on smart devices and are used by people every day. The possibility to get real-time data through off-the-shelf sensors offers to the research new challenges to face with, like «fault tolerance» problems (e.g. centralized knowledge bases vs distributed ones) that were not considered few years ago.    

This course will try to explore how the research in the expert system field is evolving, from the thoretical and practical point of view, for responding to these solicitations. 

Evaluation YES
Calendar December 2019
Advanced Distributed Systems Development with Multiagent Systems
Teacher  Dott.ssa Daniela Briola, Università degli studi di Milano Bicocca
  Prof.ssa Viviana Mascardi, Università degli studi di Genova
  Prof. Rafael Bordini, Pontificia Universidade Catolica do Rio Grande do Sul, Porto Alegre/RS - Brazil
Language Inglese
CFU  2,5
Hours  20
Program  Aims and target audience
  Multiagent systems are a general-purpose paradigm to design distributed systems, where autonomous entities collaborate to solve complex problems.
  This phd course aims at presenting the MAS paradigm from both a design and implementation point of view: the foreseen audience is a student interested in learning a paradigm for modelling a complex system, where distributed (physically but above all logically) entities/components need to cooperate to solve a problem. This kind of systems may be adopted to solve very different types of problems, so the course may help in facing complex systems in very different research areas, and is not tailored for AI students only.
  The main aspects that characterize the distributed systems that may benefit of a MAS architectural solution will be presented: the first aim of the course is to let the students begin thinking “how a MAS solution could help in this situation? How could I model it? Which platform could I use?”
  Second, we will focus on the two main MAS models and platforms, JADE and JASON (in combination with Cartago and Moise)
  •       JADE in particular will be presented in details with real MAS examples
  Educational goals:
  •    Modeling a complex system as a MAS: main approaches, problems and solutions (with real examples presented) à “designing a MAS” skill
  •    BDI paradigm (main features)
  •    Main aspects of designing an OWL ontology, and implementation with Protegè
  •    JASON modelling and programming skills, basic features
  •    A&A paradigm, with the Cartago platform
  •    Introduction to MAS modelling using a formal approach (MOISE)
  •    The JaCaMo platform
  •    JADE modelling and programming skills, basic and advances aspects
  •    JADE extensions:
           •   OntologyBeanGenerator (to integrate OWL ontologies)
           •    P2P platform (LEARN platform)
           •    Interfaces with WebServices and other languages
  MAS paradigm and JADE: 10 hrs (Dr.ssa Daniela Briola)
  •    MASs: when can they help? 
  •    Main features of MASs systems, and the FIPA standard
  •    Introduction to the JADE model for MASs
         •    Agents, Platforms, Behaviour, Messages exchange, Protocols etc
         •    Turning JADE into a basic P2P platform
  •    Introduction to the ontologies and Protegè and their integration with JADE (OntologyBeanGenerator 5)
  •    JADE in practice: solving complex problems with JADE
         •    Examples of real multiagent systems will be provided
  JASON & JACAMO: 10 hrs (Prof.ssa Viviana Mascardi, Prof. Rafael Bordini)
  •       Introduction to the BDI architecture for MASs
  •       The Agent&Artifact paradigm
  •       Introduction to JASON: main features, examples
  •       Notion about modelling a MAS with Moise and Cartago
  •       The JaCaMo platform
  JAVA is the language of both the platforms, so a basic knowledge of it is required.
  Knowledge of ontologies would avoid the part of the course regarding them, and we could move directly to the JADE integration and some more examples. TBD with the students
Evaluation Yes
Calendar February 2020
Advanced Techniques for Combinatorial Algorithms
Teacher  Gianluca Della Vedova, Raffaella Rizzi
Language English
CFU  2,5
Hours  20
Program  Parallel algorithms; Randomized algorithms; Text indexing; 
  Fixed-parameter algorithms.; Approximation algorithms;
  External-memory algorithms. Streaming algorithms  sampling)
  Text indexing 3 (Burrows-Wheeler Transform and FM-index)
Evaluation Yes
Calendar April/May 2020
Deep Learning
Teacher  Elisabetta Fersini - University of Milano-Bicocca
Language English
CFU  2.5
Hours  20
Program  -     Background concepts
  -     Training Deep Networks: 
  -     Objective functions
  -     Activation Functions
  -     Regularization
  -     Gradient-based optimization
  -     Focus on Deep Networks: 
  -     Autoencoders
  -     Convolutional Neural Networks
  -     Recurrent and Recursive Networks
  -     Practical Methodology:
  -     Performance Metrics and baseline models
  -     Selecting hyper-parameters
  -     Deep Learning in Practice: Tools, Tips and Tricks
Evaluation YES
Calendar June 2020
Do we need data structures? The top data structures you should know for next generation computing
Teacher  Paola Bonizzoni (Dominique Kempe, Solon Pissis, Kunikiko Sadakane are candidate for teaching some parts of the course)
Language English
CFU  2.5
Hours  20

The course presents  the most recent approaches to the design of data structures for dealing with the challenges of modern computing on big collections of text data. The goal of the course is the acquisition of techniques and  concepts that allow to face emerging challenges in the field of computer science.


Topics include:

      - storing and querying by succinct data structures and how to design efficient       algorithms for building such succinct data structures, 
      - Modern applications. 
      - Data streaming and algorithms for dealing with text data
Evaluation YES
Calendar March-April 2020
Neural Symbolic Computation
Teacher  Guido Fiorino – Università degli Studi di Milano-Bicocca
  Italo Zoppis – Università degli Studi di Milano-Bicocca
  Rafael Peñaloza – Università degli Studi di Milano-Bicocca
  Luciano Serafini – Fondazione Bruno Kessler
Language English
CFU  2.5
Hours  20
Program  Logics for Knowledge Representation
  Non-classical Logics for Neural-Symbolic Integration
  Neural Network Architectures for Neural-Symbolic Integration
  Graph Neural Networks and Fibring Neural Networks
  Neural-Symbolic Learning Systems and CILP
  Connectionist Languages
  Relational Learning and Uncertainty Reasoning
  Introduction to Logic Tensor Networks
  Hands-on Learning and Reasoning with LTNs
Evaluation YES
Calendar September 2020
a cura di Scuola di dottorato, ultimo aggiornamento il 24/10/2019