Node-to-Neighborhood Semantic Consistency: Text-Topology Alignment for TAGs Anomaly Detection
2026-06-29 • Computation and Language
Computation and Language
AI summaryⓘ
The authors focus on finding unusual or suspicious nodes in graphs where each node has related text, like in fraud detection. They point out that current methods either understand the graph’s structure well or the text well, but not both together, and they miss checking if a node’s text matches its neighbors’ context. To fix this, the authors propose a method called N2NSC that looks at how a node’s text and its neighboring nodes’ structure relate to each other. Their approach combines both text and graph information and does better than previous methods on multiple datasets.
graph anomaly detectiontext-attributed graphsgraph neural networkslarge language modelssemantic consistencygraph topologynode-to-neighborhood relationshipfraud detectionnode anomaly
Authors
Bochen Lin, Jianxiang Yu, Jiayi Wu, Lin Qi, Huang Lu, Xiang Li
Abstract
Graph anomaly detection (GAD) on text-attributed graphs (TAGs) is vital for applications such as fraud detection and academic integrity verification. Existing approaches generally fall into two paradigms. GNN-based methods effectively capture structural patterns but struggle to capture fine-grained textual semantics. Methods integrating LLMs with graphs improve semantic understanding yet fail to fully comprehend topological relationships among neighboring nodes. Moreover, both paradigms overlook the correspondence between textual semantics and graph topological relationships, limiting their ability to identify nodes whose semantics are inconsistent with their neighborhoods. In this paper, we formalize TAG anomaly detection as a node-to-neighborhood semantic consistency problem, where anomalies may arise from either textual semantic mismatch or topological deviation between a node and its neighbors. We propose N2NSC (Node-to-Neighborhood Semantic Consistency), a framework that captures the correspondence between graph topology and textual semantics through two complementary fusion paths. The two pathways work synergistically, enabling the LLM to fully leverage both textual and structural neighborhood information for anomaly detection. Extensive experiments across eight datasets demonstrate that N2NSC consistently outperforms current state-of-the-art methods.