Hedged learning: regret-minimization with learning experts

In non-cooperative multi-agent situations, there cannot exist a globally optimal, yet opponent-independent learning algorithm. Regret-minimization over a set of strategies optimized for potential opponent models is proposed as a good framework for deciding how to behave in such situations. Using longer playing horizons and experts that learn as they play, the regret-minimization framework can be extended to overcome several shortcomings of earlier approaches to the problem of multi-agent learning.

• Type: Conference
• Year: 2005
• Where: ICML
• Authors: Yu-Han Chang, Leslie Pack Kaelbling
• Keywords: ICML 2005, Machine Learning, Multi-agent Learning, Non-cooperative Multi-agent Situations, Potential Opponent Models