ICML
2005
IEEE
14 years 5 months ago
2005
IEEE
In non-cooperative multi-agent situations, there cannot exist a globally optimal, yet opponent-independent learning algorithm. Regret-minimization over a set of strategies optimiz...