Search

Few direct matches — filled in with the latest updates.

Tag: #co-player-inference1 results

In-Context Inference Enables Multi-Agent Cooperation

In-Context Inference Enables Multi-Agent Cooperation

Researchers demonstrate that sequence models' in-context learning induces cooperation in multi-agent RL without hardcoded co-player assumptions or timescale separation. Training against diverse co-players leads to best-response strategies on intra-episode timescales. This naturally emerges mutual shaping via extortion vulnerability, providing a scalable path to cooperative behaviors.