SGOOD: Substructure-enhanced Graph-Level Out-of-Distribution Detection
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2023
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866911959844126720 |
|---|---|
| author | Ding, Zhihao Shi, Jieming Shen, Shiqi Shang, Xuequn Cao, Jiannong Wang, Zhipeng Gong, Zhi |
| author_facet | Ding, Zhihao Shi, Jieming Shen, Shiqi Shang, Xuequn Cao, Jiannong Wang, Zhipeng Gong, Zhi |
| contents | Graph-level representation learning is important in a wide range of applications. Existing graph-level models are generally built on i.i.d. assumption for both training and testing graphs. However, in an open world, models can encounter out-of-distribution (OOD) testing graphs that are from different distributions unknown during training. A trustworthy model should be able to detect OOD graphs to avoid unreliable predictions, while producing accurate in-distribution (ID) predictions. To achieve this, we present SGOOD, a novel graph-level OOD detection framework. We find that substructure differences commonly exist between ID and OOD graphs, and design SGOOD with a series of techniques to encode task-agnostic substructures for effective OOD detection. Specifically, we build a super graph of substructures for every graph, and develop a two-level graph encoding pipeline that works on both original graphs and super graphs to obtain substructure-enhanced graph representations. We then devise substructure-preserving graph augmentation techniques to further capture more substructure semantics of ID graphs. Extensive experiments against 11 competitors on numerous graph datasets demonstrate the superiority of SGOOD, often surpassing existing methods by a significant margin. The code is available at https://github.com/TommyDzh/SGOOD. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2310_10237 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | SGOOD: Substructure-enhanced Graph-Level Out-of-Distribution Detection Ding, Zhihao Shi, Jieming Shen, Shiqi Shang, Xuequn Cao, Jiannong Wang, Zhipeng Gong, Zhi Machine Learning Artificial Intelligence Graph-level representation learning is important in a wide range of applications. Existing graph-level models are generally built on i.i.d. assumption for both training and testing graphs. However, in an open world, models can encounter out-of-distribution (OOD) testing graphs that are from different distributions unknown during training. A trustworthy model should be able to detect OOD graphs to avoid unreliable predictions, while producing accurate in-distribution (ID) predictions. To achieve this, we present SGOOD, a novel graph-level OOD detection framework. We find that substructure differences commonly exist between ID and OOD graphs, and design SGOOD with a series of techniques to encode task-agnostic substructures for effective OOD detection. Specifically, we build a super graph of substructures for every graph, and develop a two-level graph encoding pipeline that works on both original graphs and super graphs to obtain substructure-enhanced graph representations. We then devise substructure-preserving graph augmentation techniques to further capture more substructure semantics of ID graphs. Extensive experiments against 11 competitors on numerous graph datasets demonstrate the superiority of SGOOD, often surpassing existing methods by a significant margin. The code is available at https://github.com/TommyDzh/SGOOD. |
| title | SGOOD: Substructure-enhanced Graph-Level Out-of-Distribution Detection |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2310.10237 |