SMAX

The 'SMAX' dataset/benchmark is part of a diagnostic suite used to evaluate the behavioral complexity of baseline policies in cooperative multi-agent reinforcement learning scenarios, specifically assessing the necessity of Dec-POMDP reasoning for success.

Papers using SMAX (3)

Quality-Aware Exploration Budget Allocation for Cooperative Multi-Agent Reinforcement Learning2026Probing Dec-POMDP Reasoning in Cooperative MARL2026Simplifying Deep Temporal Difference Learning2024 · 1 cites