Games, Dynamics, and Learning Minicourse
Section outline
-
This mini-course is intended to provide a short introduction to the subject of learning in games – from models of evolution in population games to online learning and regret minimization. The course will consist of three main parts, depending on the type of games considered – nonatomic, finite, or continuous. In particular, we will cover the basics of evolutionary dynamics in population games (insisting on the replicator equation and its rationality properties), no-regret learning in finite games (focusing on the multiplicative/exponential weights algorithm and its properties), and the basics of online optimization in continuous games (online gradient descent, follow the regularizer leader, etc.). We will also discuss the impact of the information available to the players, as well as a range of concrete applications to machine learning and operations research.Every Friday 16:00-19:00Lectures by Panayiotis MertikopoulosPasscode: 54a8SP
-
Background Material• Introduction
• Nash equilibrium
• Other notions of rationality (dominated strategies, correlated equilibrium,...)
-
Game Dynamics
• Introduction
• The replicator dynamics
• Rationality analysis
• Other topics (imitative dynamics, best-reply dynamics,...) -
Learning in finite games
• Introduction
• No-regret learning in cont. time
• No-regret learning in discrete time
• Other topics (fictitious play and its variants,...)
-
Learning in continuous games
• Introduction
• No-regret learning in cont. time
• No-regret learning in discrete time
• Other topics (optimistic algorithms,...)