RESCORE: LLM-Driven Simulation Recovery in Control Systems Research Papers

arXiv 2026 47.3 method

TLDR

RESCORE uses LLM agents to recover simulations from control systems papers, achieving 40.7% success and 10x speedup over manual replication.

Reasoning

The paper introduces a novel task and benchmark for simulation recovery, with a clear methodology and empirical results. However, the success rate is modest (40.7%) and the approach is domain-specific to control systems, limiting generalizability.

Read-first score

Read-first score 47.3, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 36.

Recency 8%

100

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

Methodology quality 25%

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

Reproducibility 25%

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

Topical relevance 42%

Uses existing LLM keyword relevance scores normalized to 0-100. AI scientist,automated scientific discovery,autonomous research agent,automated research,literature review agent,survey generation,automated experimentation,experiment design agent,AI for scientific research,paper writing agent,research automation,scientific discovery agent

Field roles

Frontier

Rank sensitivity

Stability: volatile; rank range: 28.

Keyword Scores

research automation

automated research

AI for scientific research

automated scientific discovery

autonomous research agent

scientific discovery agent

AI scientist

automated experimentation

literature review agent

survey generation

experiment design agent

paper writing agent

Deep Analysis

Innovations

Defines the task of Paper to Simulation Recoverability for control systems research papers.
Curates a benchmark of 500 papers from IEEE Conference on Decision and Control (CDC).
Proposes RESCORE, a three-component LLM agentic framework (Analyzer, Coder, Verifier) with iterative execution feedback and visual comparison.

Methodology

RESCORE is an LLM agentic framework with Analyzer, Coder, and Verifier components that uses iterative execution feedback and visual comparison to generate executable simulation code from control systems papers. It is evaluated on a curated benchmark of 500 CDC papers.

Key Results

RESCORE recovers task coherent simulations for 40.7% of benchmark instances, outperforming single pass generation, and achieves an estimated 10X speedup over manual human replication.

Resources

Paper