ICML
2005
IEEE
15 years 13 days 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...