Court of LLMs: Evidence-Augmented Generation via Multi-LLM Collaboration for Text-Attributed Graph Anomaly Detection

2025-08-01

AI Summary

Existing text-attributed graph (TAG) anomaly detection methods predominantly rely on shallow text encoders, leading to insufficient semantic contextual modeling. Although large language models (LLMs) offer strong semantic understanding, their fixed input-length constraints hinder effective modeling of higher-order graph structures. To address this, we propose an LLM-GNN collaborative framework: for the first time, we employ multiple LLMs in concert to generate structure-aware semantic evidence; this is integrated with a graph neural network (GNN) to capture topological dependencies and aligned via a gated fusion mechanism that bridges structural and semantic representations. Evaluated on multiple benchmarks, our method achieves an average 13.37% improvement in Average Precision (AP) over state-of-the-art approaches. It delivers both high detection accuracy and model interpretability, empirically validating the effectiveness and necessity of jointly modeling textual semantics and graph structure.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Sentence-level Semantics, Textual Inference, etc.
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information

Abstract

The natural combination of intricate topological structures and rich textual information in text-attributed graphs (TAGs) opens up a novel perspective for graph anomaly detection (GAD). However, existing GAD methods primarily focus on designing complex optimization objectives within the graph domain, overlooking the complementary value of the textual modality, whose features are often encoded by shallow embedding techniques, such as bag-of-words or skip-gram, so that semantic context related to anomalies may be missed. To unleash the enormous potential of textual modality, large language models (LLMs) have emerged as promising alternatives due to their strong semantic understanding and reasoning capabilities. Nevertheless, their application to TAG anomaly detection remains nascent, and they struggle to encode high-order structural information inherent in graphs due to input length constraints. For high-quality anomaly detection in TAGs, we propose CoLL, a novel framework that combines LLMs and graph neural networks (GNNs) to leverage their complementary strengths. CoLL employs multi-LLM collaboration for evidence-augmented generation to capture anomaly-relevant contexts while delivering human-readable rationales for detected anomalies. Moreover, CoLL integrates a GNN equipped with a gating mechanism to adaptively fuse textual features with evidence while preserving high-order topological information. Extensive experiments demonstrate the superiority of CoLL, achieving an average improvement of 13.37% in AP. This study opens a new avenue for incorporating LLMs in advancing GAD.

Problem

• Detecting anomalies in text-attributed graphs using multi-modal data
• Overcoming limitations of shallow text encoding in graph anomaly detection
• Integrating LLMs and GNNs for improved semantic and structural analysis

Innovation

• Multi-LLM collaboration for evidence-augmented generation
• GNN with gating mechanism for adaptive fusion
• Combining LLMs and GNNs for anomaly detection