Reproduce “Error-Propagation Modeling for Failure Attribution in LLM-Based Multi-Agent Systems” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Jiaqi Liao, Yuanzhao Zhai, Huanxi Liu, Xu Zhang, Zheming Zhuang, Dawei Feng, Bo Ding, Huaimin Wang
EMFA diagnoses decisive errors in failed multi-agent workflows by tracing dependencies, identifying persistent unproductive loops, and testing candidate repairs with model-based counterfactual assessment. On the Who&When benchmark, it achieves 50.00% exact-step accuracy versus CHIEF’s 45.60% on Algorithm-Generated logs and leads on step accuracy across both subsets. The evidence is limited to 58 Hand-Crafted and 126 Algorithm-Generated logs.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.