AD-AGENT: A Multi-agent Framework for End-to-end Anomaly Detection

2025-05-19

AI Summary

Non-expert users face significant challenges in performing end-to-end anomaly detection (AD) across multimodal data and heterogeneous AD libraries (e.g., PyOD, PyGOD, TSLib). Method: We propose the first LLM-driven multi-agent collaborative framework that automatically compiles natural-language instructions into executable AD pipelines. It employs intent parsing, documentation-aware iterative code generation, and debugging within a shared short-term workspace and long-term cache—requiring no user programming expertise or domain knowledge. The framework supports cross-library model invocation and multimodal input handling. Contribution/Results: Experimental evaluation demonstrates high reliability of generated scripts and AD model recommendations matching expert-level performance. The system is open-sourced, substantially lowering the barrier to practical AD deployment and accelerating both real-world adoption and research translation.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Multiagent Systems: Adversarial Agents
• Data Mining & Knowledge Management: Anomaly/Outlier Detection

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Anomaly detection (AD) is essential in areas such as fraud detection, network monitoring, and scientific research. However, the diversity of data modalities and the increasing number of specialized AD libraries pose challenges for non-expert users who lack in-depth library-specific knowledge and advanced programming skills. To tackle this, we present AD-AGENT, an LLM-driven multi-agent framework that turns natural-language instructions into fully executable AD pipelines. AD-AGENT coordinates specialized agents for intent parsing, data preparation, library and model selection, documentation mining, and iterative code generation and debugging. Using a shared short-term workspace and a long-term cache, the agents integrate popular AD libraries like PyOD, PyGOD, and TSLib into a unified workflow. Experiments demonstrate that AD-AGENT produces reliable scripts and recommends competitive models across libraries. The system is open-sourced to support further research and practical applications in AD.

Problem

• Addresses challenges in anomaly detection for non-expert users
• Integrates diverse AD libraries into a unified workflow
• Generates executable pipelines from natural-language instructions

Innovation

• LLM-driven multi-agent framework for AD
• Natural-language to executable AD pipelines
• Integrates popular AD libraries into workflow