GUARDIAN: Safeguarding LLM Multi-Agent Collaborations with Temporal Graph Modeling
This addresses safety vulnerabilities in LLM-based multi-agent systems, which is crucial for reliable AI applications, though it appears incremental as it builds on existing graph modeling and anomaly detection techniques.
The paper tackles safety challenges like hallucination amplification and error propagation in multi-agent LLM collaborations by introducing GUARDIAN, a method that models interactions as temporal graphs and uses an unsupervised encoder-decoder to detect anomalies, achieving state-of-the-art accuracy.
The emergence of large language models (LLMs) enables the development of intelligent agents capable of engaging in complex and multi-turn dialogues. However, multi-agent collaboration faces critical safety challenges, such as hallucination amplification and error injection and propagation. This paper presents GUARDIAN, a unified method for detecting and mitigating multiple safety concerns in GUARDing Intelligent Agent collaboratioNs. By modeling the multi-agent collaboration process as a discrete-time temporal attributed graph, GUARDIAN explicitly captures the propagation dynamics of hallucinations and errors. The unsupervised encoder-decoder architecture incorporating an incremental training paradigm learns to reconstruct node attributes and graph structures from latent embeddings, enabling the identification of anomalous nodes and edges with unparalleled precision. Moreover, we introduce a graph abstraction mechanism based on the Information Bottleneck Theory, which compresses temporal interaction graphs while preserving essential patterns. Extensive experiments demonstrate GUARDIAN's effectiveness in safeguarding LLM multi-agent collaborations against diverse safety vulnerabilities, achieving state-of-the-art accuracy with efficient resource utilization. The code is available at https://github.com/JialongZhou666/GUARDIAN