XEP36AGT – Advanced Computational Game Theory

This is a CW page for the course Advanced Computational Game Theory (XEP36AGT). The course cover 3 areas from computational and algorithmic game theory:

  1. equilibrium computation in finite games: complexity classes, standard exact and approximate algorithms for finding various game-theoretic solution concepts
  2. reinforcement learning methods in games
  3. solving continuous games

The course will be a combination of lectures on a given topic (80%) and reading group (20%) where students will study in details recent papers of their choice that is closely related to the topics covered during the lectures and will present their analysis.

The lectures are going to take place on Friday, 12:30, in KN:E-205 (first lecture will be 25.2.2022).

Lectures

Date Topic Lecturer Current Slides Old Slides
Introduction, Overview of Equilibria, Main Complexity Classes, Nash, Fixed Point Bosansky l01_2019.pdf 02_2019.pdf 02.pdf
Computing and Approximating a Nash Equilibrium (Lemke Howson, MILP) Bosansky 03_2019.pdf 03.pdf04.pdf
Computing a Stackelberg Equilibrium Bosansky 05_2019.pdf 05.pdf
Computing and Approximation of a Correlated Equilibrium Bosansky 06_2019.pdf 06.pdf
Correlated Equilibrium in Succinct Games, Other Dynamic Games Bosansky 07_2019.pdf 07.pdf
Online Learning and Multiarmed Bandit Problems Lisy agt_learning1_2019.pdf agt_learning1_2018.pdf
Learning in Normal-Form Games, Fictitious Play Lisy agt_learning2_2019.pptx agt_learning2_2018.pdf
Regret Matching, Counterfactual Regret Minimization Lisy agt_learning3_2019.pdf agt_learning3_2018.pdf
Continual Resolving in Extensive-Form Games (DeepStack) Lisy agt_learning4_2019.pdf agt_learning4_2018.pdf
Continuous Games and Their Equilibria. Separable Games. Kroupa kroupa-cg1_1_.pdf
Polynomial Games. Reduction to an SDP Problem. Kroupa kroupa-cg2.pdf

Class projects

courses/xep36agt/start.txt · Last modified: 2022/02/24 20:31 by bosanbra