{"type":"ScholarlyArticle","title":"Unsupervised Graph-Level Anomaly Detection via Multi-granular Graph Structure Learning","authors":[{"name":"Ge Zhang","slug":"ge-zhang","given":"Ge","family":"Zhang","affiliation":""},{"name":"Huimei Li","slug":"huimei-li","given":"Huimei","family":"Li","affiliation":""},{"name":"Guohao Sun","slug":"guohao-sun","given":"Guohao","family":"Sun","affiliation":""},{"name":"Xiu Fang","slug":"xiu-fang","given":"Xiu","family":"Fang","affiliation":""},{"name":"Xixun Lin","slug":"xixun-lin","given":"Xixun","family":"Lin","affiliation":""},{"name":"Xiaobao Wang","slug":"xiaobao-wang","given":"Xiaobao","family":"Wang","affiliation":""},{"name":"Pengfei Jiao","slug":"pengfei-jiao","given":"Pengfei","family":"Jiao","affiliation":""},{"name":"Liang Yang","slug":"liang-yang","given":"Liang","family":"Yang","affiliation":""}],"abstract":"Graph-level anomaly detection (GLAD) aims to identify graphs that deviate from the majority in a dataset of graphs. Existing methods typically adopt either a global aggregation perspective that summarizes nodes within a graph into a representation vector, or a subgraph-oriented perspective which regards certain local subgraphs as indicators of anomalous properties. However, both paradigms operate from a fixed perspective and may fail to identify anomalous graphs whose discriminative characteristics manifest at multiple levels of structure granularity, where each level features the coarsened graphs at a specific granularity. In this paper, we propose M-GLAD, an unsupervised GLAD method via multi-granular graph structure learning. M-GLAD is grounded in the Prototype-guided Multi-granular Information Bottleneck (PMIB) principle. PMIB aims to measure the mutual information between coarsened graphs at the specific level of structure granularity and the learnable granularity-specific prototypes that summarize normal patterns of graphs at each granularity. The multiple levels of structure granularity of graphs and prototypes are abstracted by a graph structure abstraction module. This formulation enables granularity-aware anomaly scoring that considers anomalous characteristics across different levels of structure granularity. Extensive experiments on eight real-world graph datasets demonstrate that M-GLAD achieves superior performance over competitive baselines.","doi":"10.24963/ijcai.2026/355","url":"https://ijcai.org/proceedings/2026/355","pdf":"https://ijcai.org/proceedings/2026/0355.pdf","markdown":"https://ijcai.org/proceedings/2026/355.md","bibtex":"https://ijcai.org/proceedings/2026/bibtex/355","volume":{"slug":"2026","title":"Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence","short_title":"IJCAI 2026","url":"https://ijcai.org/proceedings/2026/","year":2026,"location":"Bremen, Germany","dates":"15-21 August 2026","editors":"Diego Calvanese","isbn":"978-1-956792-09-6","publisher":"International Joint Conferences on Artificial Intelligence Organization","doi_prefix":"10.24963/ijcai.2026"},"section":"Data Mining","track":"Main Track","number":355,"pages":{"first":3196,"last":3204},"keywords":["Data Mining: Anomaly/outlier detection","Data Mining: Mining graphs"],"videos":[],"references":[],"license_note":"Copyright International Joint Conferences on Artificial Intelligence."}