Reproduce “FaultExplainer: Leveraging Large Language Models for Interpretable Fault Detection and Diagnosis” 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
Abdullah Khan, Rahul Nahar, Hao Chen, Gonzalo E. Constante Flores, Can Li
Purdue University researchers developed FaultExplainer, a system integrating statistical fault detection with large language models to provide interpretable, natural language explanations for chemical process faults. The system accurately diagnosed 9 of 11 detectable faults using known root causes and 8 of 11 unseen faults, demonstrating grounded reasoning by linking explanations to objective process data.
Key insights
- Combining traditional statistical FDD with LLMs effectively bridges the gap between quantitative data analysis and intuitive, natural language explanations for process faults.
- LLMs can successfully infer and explain potential root causes for previously unseen faults when adequately prompted and provided with relevant process context and feature deviations.
- Explicitly grounding LLM outputs with verifiable process data and domain knowledge is crucial for ensuring accuracy and trustworthiness in safety-critical applications like FDD.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.