conftrace_

2020 ACML ACML 2020

Foolproof Cooperative Learning

Abstract

This paper extends the notion of learning algorithms and learning equilibriums from repeated games theory to stochastic games. We introduce Foolproof Cooperative Learning (FCL), an algorithm that converges to an equilibrium strategy that allows cooperative strategies in self-play setting while being not exploitable by selfish learners. By construction, FCL is a learning equilibrium for repeated symmetric games. We illustrate the behavior of FCL on symmetric matrix and grid games, and its robustness to selfish learners.

Authors

Alexis Jacq, Julien Pérolat, Matthieu Geist, Olivier Pietquin

Topics

Artificial Intelligence > Core AI > Multi-Agent Systems Machine Learning > Learning Types > Adversarial Learning Machine Learning > Optimization & Theory > Learning Theory Artificial Intelligence > Core AI > Game Theory Machine Learning > Optimization & Theory > Game Theory

Keywords

game theory equilibrium learning self-play learning repeated game stochastic game cooperative learning equilibrium strategy multi-agent system cooperative strategy learning equilibrium

More from ACML 2020

All ACML 2020 papers →

Enhancing Topic Models by Incorporating Explicit and Implicit External Knowledge

Exact Passive-Aggressive Algorithms for Multiclass Classification Using Bandit Feedbacks

FIREPruning: Learning-based Filter Pruning for Convolutional Neural Network Compression

Geodesically-convex optimization for averaging partially observed covariance matrices

Inferring Continuous Treatment Doses from Historical Data via Model-Based Entropy-Regularized Reinforcement Learning

Learning 2-opt Heuristics for the Traveling Salesman Problem via Deep Reinforcement Learning