PolicyEvolve evolves LLM-generated rule-based policies via a global and local pool plus trajectory reflection, and reports winning head-to-head comparisons in the Wrestle game.
Google research football: A novel reinforcement learning environment
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
PolicyEvolve: Evolving Programmatic Policies by LLMs for multi-player games via Population-Based Training
PolicyEvolve evolves LLM-generated rule-based policies via a global and local pool plus trajectory reflection, and reports winning head-to-head comparisons in the Wrestle game.