GUARDIAN: Safeguarding LLM Multi-Agent Collaborations with Temporal Graph Modeling

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhou, Jialong, Wang, Lichao, Yang, Xiao
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909848036179968
author Zhou, Jialong
Wang, Lichao
Yang, Xiao
author_facet Zhou, Jialong
Wang, Lichao
Yang, Xiao
contents 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
format Preprint
id arxiv_https___arxiv_org_abs_2505_19234
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle GUARDIAN: Safeguarding LLM Multi-Agent Collaborations with Temporal Graph Modeling
Zhou, Jialong
Wang, Lichao
Yang, Xiao
Artificial Intelligence
Computation and Language
Multiagent Systems
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
title GUARDIAN: Safeguarding LLM Multi-Agent Collaborations with Temporal Graph Modeling
topic Artificial Intelligence
Computation and Language
Multiagent Systems
url https://arxiv.org/abs/2505.19234