Awesome World Model Hub Papers · Datasets · Projects
← Back to papers

PatchWorld: Gradient-Free Optimization of Executable World Models

arXiv 2026 60.3 method

TLDR

PatchWorld induces executable Python world models from offline trajectories via counterexample-guided code repair, achieving high planning scores without LLM calls.

Reasoning

The paper introduces a novel gradient-free method for creating symbolic, inspectable world models from offline data, with strong empirical results on AgentGym environments. However, the approach is limited to text-agent domains and reveals a tradeoff between observation fidelity and decision utility that may constrain practical use.

Read-first score

Read-first score 60.3, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 35.

Recency 6%
100

Uses a gentle age decay so recent papers surface without erasing older foundations. 2026

Citation impact 18%
84

Uses OpenAlex-shaped citation metadata as a bibliometric attention signal, separate from paper quality. citation_normalized_percentile=0.83957915

Reproducibility 18%
81

Screens links and visible text for paper, code, dataset, artifact, and repository signals. pdf=True; code=True; dataset=False; markers=code,github

Methodology quality 18%
60

Screens visible abstract and analysis fields for experiment, dataset, baseline, metric, and limitation evidence. markers=baseline

Topical relevance 29%
50

Uses existing LLM keyword relevance scores normalized to 0-100. world model,world simulator,generative world model,interactive world model,video world model,world dynamics prediction,model-based reinforcement learning world model

Citation velocity 12%
0

Citation velocity estimates citations per publication-year to reduce old-paper bias. velocity=0.00

Field roles

FoundationFrontierBridgeReproducibility anchor

Rank sensitivity

Stability: volatile; rank range: 462.

Keyword Scores

world model
10
world simulator
7
world dynamics prediction
7
interactive world model
5
model-based reinforcement learning world model
4
generative world model
2
video world model
0

Deep Analysis

Innovations

  • Gradient-free framework that turns offline trajectories into executable Python world models via counterexample-guided code repair
  • Induces symbolic belief-state programs whose action updates can be inspected, replayed, and locally patched
  • No LLM calls inside the world-model prediction module itself

Methodology

PatchWorld uses offline trajectories to induce executable Python world models through counterexample-guided code repair. It produces symbolic belief-state programs that can be inspected and patched locally. The method is evaluated on seven AgentGym environments using code-based planning scores and compared against baselines.

Key Results

PatchWorld-Simple achieves the highest code-based planning score among evaluated methods, reaching 76.4% macro success in live one-step lookahead while invoking no LLM calls inside the world-model prediction module.

Limitations

  • A human-specified residual-memory bias improves surface observation fidelity but weakens decision utility
  • Improving observation fidelity can come at the expense of action-discriminative dynamics, exposing a tradeoff in executable world models

Tags

gradient-free optimizationworld modelsexecutable codecounterexample-guided repairplanningCLAI