AutoIAD: Manager-Driven Multi-Agent Collaboration for Automated Industrial Anomaly Detection

2025-08-07

AI Summary

Industrial Anomaly Detection (IAD) suffers from high scene-adaptation costs and heavy reliance on manual hyperparameter tuning. To address this, we propose the first end-to-end multi-agent automation framework tailored for IAD: a manager agent orchestrates specialized sub-agents—including data preprocessing, model architecture design, and training agents—while tightly integrating an industrial anomaly domain knowledge base. Leveraging large language models (LLMs), our modular agent architecture employs iterative optimization and knowledge-guided reasoning to fully automate the pipeline from raw images to high-performance anomaly detection models. On the MVTec AD benchmark, our method significantly outperforms general-purpose agent and AutoML baselines, achieving substantial AUROC gains. Ablation studies confirm that the manager agent mechanism and the domain knowledge base are critical for improving task completion rate, model robustness, and hallucination suppression.

Technology Category

• Multiagent Systems: Adversarial Agents
• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Data Mining & Knowledge Management: Anomaly/Outlier Detection

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Agentic search

Abstract

Industrial anomaly detection (IAD) is critical for manufacturing quality control, but conventionally requires significant manual effort for various application scenarios. This paper introduces AutoIAD, a multi-agent collaboration framework, specifically designed for end-to-end automated development of industrial visual anomaly detection. AutoIAD leverages a Manager-Driven central agent to orchestrate specialized sub-agents (including Data Preparation, Data Loader, Model Designer, Trainer) and integrates a domain-specific knowledge base, which intelligently handles the entire pipeline using raw industrial image data to develop a trained anomaly detection model. We construct a comprehensive benchmark using MVTec AD datasets to evaluate AutoIAD across various LLM backends. Extensive experiments demonstrate that AutoIAD significantly outperforms existing general-purpose agentic collaboration frameworks and traditional AutoML frameworks in task completion rate and model performance (AUROC), while effectively mitigating issues like hallucination through iterative refinement. Ablation studies further confirm the crucial roles of the Manager central agent and the domain knowledge base module in producing robust and high-quality IAD solutions.

Problem

• Automates industrial anomaly detection pipeline development
• Reduces manual effort in manufacturing quality control
• Improves model performance and task completion rates

Innovation

• Manager-Driven multi-agent collaboration framework
• Domain-specific knowledge base integration
• Automated end-to-end anomaly detection pipeline