PatchWorld induces executable Python belief-state programs from offline trajectories via gradient-free counterexample-guided code repair, achieving 76.4% macro success in one-step lookahead planning across seven AgentGym environments without LLM calls in the prediction module.
Armando Solar-Lezama, Liviu Tancau, Rastislav Bodik, Sanjit Seshia, and Vijay Saraswat
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
PatchWorld: Gradient-Free Optimization of Executable World Models
PatchWorld induces executable Python belief-state programs from offline trajectories via gradient-free counterexample-guided code repair, achieving 76.4% macro success in one-step lookahead planning across seven AgentGym environments without LLM calls in the prediction module.