This is the GitHub for the Theory module of the Decision Support Systems course, offered at the 1st year of the Master of Science in Artificial Intelligence at University of Milano-Bicocca.
The aim of the course is to provide an intensive, theoretically-oriented introduction to normative decision theory (incl. single-agent decision theory, non-cooperative game theory, coalitional game theory and social choice theory), as well as applications in modern Artificial Intelligence (specifically, Machine Learning). The course adopts a modern teaching approach, wherein traditional "basic" content is tightly integrated with research-oriented lessons inspired by current developments in the field.
Exam mode: Written exam + Oral exam (optional)
The written exam will be an open question assessment on the core contents taught in lesson. Questions will regard either basic definitions and notions
or hybrid exercise/theoretical questions.
Oral examination consists of preparation and discussion of an in-depth presentation on the contents of a selected paper (either among those proposed by
the lecturer or also proposed by the student, after agreement with the lecturer) related to the topics of the course.
Instructor Name: Andrea Campagner (PhD)
Instructor Affiliation: Post-Doc Researcher at IRCCS Istituto Ortopedico Galeazzi, Milan, Italy
Reference Material:
- Slides prepared by the lecturer
- Selected articles (shared with the class prior to the lessons)
- Y. Shoham, K. Leyton-Brown (2009). Multiagent Systems: Algorithmic, Game-Theoretic, and Logical Foundations. Cambridge University Press
- M. Peterson (2017). An Introduction to Decision Theory (2nd ed.). Cambridge University Press
- F. Brandt, V. Conitzer, U. Endriss, J. Lang, A. Procaccia (Eds.)(2016). Handbook of Computational Social Choice. Cambridge University Press
Lesson Plan (12h + 1h)
- 12/04/2023 - Introduction to Decision Theory, Normative vs Descriptive approaches, Decision under Ignorance, Decision under Risk (2h)
- 18/04/2023 - Utility and Decision Theory in ML: Cost-sensitive Classification, Net Benefit and Decision Curves (1h)
- 18/04/2023 - Non-cooperative Game Theory: basic concepts, Pareto optimality, Nash Equilibria and Nash Theorem, Minimax Theorem, Computation and Hints to Complexity (3h)
- 19/04/2023 - Applications of Game Theory in ML: Generative Adversarial Network (2h)
- 02/05/2023 - Seminar: Prof. Piotr Artiemjew, University of Warmia and Mazury in Olsztyn (Olsztyn, Poland). Granularity, Uncertainty and Aggregation in ML (1h)
- 03/05/2023 - Coalitional Game Theory: basic concepts, Shapley values, core, nucleolus, theoretical characterizations and computation (2h)
- 09/05/2023 - Social choice theory: basic concepts, canonical examples of social choice functions, basic theoretical results and interpretation, strategyproofness and manipulability (1h)
- 30/05/2023 - Theoretical and hands-on introduction to the DSS Quality Assessment tool (https://dss-quality-assessment.vercel.app/) (1h)