The Algorithmic Learning in Games seminar is a joint effort between faculty at London School of Economics, the University of Waterloo, Imperial College London and, King's College London.
We meet once every fortnight to discuss problems and research in a range of topics related to:
Game Theory
Multi-Agent Reinforcement Learning
Learning in Games
Multi-Agent Systems
Recordings of our past sessions can be found here.
Next Talk
Our next talk will be given by Daniele Condorelli on January 20th 2025 at 16:00 GMT. Please see further details below:
Deep Learning to Play Games
Daniele Condorelli - Department of Economics, University of Warwick
We train two neural networks adversarially to play normal-form games. At each iteration, a row and column network take a new randomly generated game and output individual mixed strategies. The parameters of each network are independently updated via stochastic gradient descent to minimize expected regret given the opponent’s strategy. Our simulations demonstrate that the joint behavior of the networks converges to strategies close to Nash equilibria in almost all games. For all 2 ×2 and in 80% of 3 ×3 games with multiple equilibria, the networks select the risk-dominant equilibrium. Our results show how Nash equilibrium emerges from learning across heterogeneous games.
Organisers
London School of Economics
University of Waterloo
Imperial College London
King's College London
Schedule
We intend to meet every other Monday from 16:00 - 17:00 GMT - please allow for slight changes to this schedule as needed.
Upcoming Speakers
Past Speakers
Recordings of past sessions can be found here or via the title of the respective talk in the spreadsheet below.
Sign-Up Form
We are open to any interested in joining the seminar, both as a member of the audience and as a speaker. To get in touch, please fill out the form below and we will get back to you as soon as possible.