Research map / Graph learning / Graph self-supervised learning
Spatial-Temporal Graph Learning with Adversarial Contrastive Adaptation
Qianru Zhang, Chao Huang, Lianghao Xia, Zheng Wang, Siu Ming Yiu, Ruihua Han
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Adding adversarial contrastive view (ADS) to the model (ablation) | GraphST · NYC-Taxi and NYC-Bike traffic | MAE: NYC-Taxi 1.3731 -> 1.2892; NYC-Bike 0.9067 -> 0.7787 | Improved |
| Adding information minimization (IM) (ablation) | GraphST · NYC-Taxi and NYC-Bike traffic | MAE: NYC-Taxi 1.3519 -> 1.2892; NYC-Bike 0.9132 -> 0.7787 | Improved |
| Replacing random edge dropout with VGAE-based adaptive augmentation (adding VGAE) (ablation) | GraphST · NYC-Taxi and NYC-Bike traffic | MAE: NYC-Taxi 1.3486 -> 1.2892; NYC-Bike 0.8941 -> 0.7787 | Improved |
| Adding information regularization (IR) (ablation) | GraphST · NYC-Taxi and NYC-Bike traffic | MAE: NYC-Taxi 1.3418 -> 1.2892; NYC-Bike 0.8837 -> 0.7787 | Improved |
| Full GraphST vs. strongest baselines across crime, traffic, and house price tasks | GraphST · Multiple (Table 2) | MAE/MAPE/RMSE: e.g., NYC-Crime MAE 2.1060 vs best baseline MGFN 3.5971 | Improved |
| Pre-training GraphST embeddings then fine-tuning with ST-SHN for crime forecasting | ST-SHN · Chicago and NYC crime (per-category, Table 4) | MAE: Chicago Theft 0.9107 vs MGFN 0.9231 | Improved |
| Replacing VGAE with random edge dropout (w/o VGAE) combined with other ablations | GraphST · Sparse data robustness | qualitative | No clear effect |
Closest papers by meaning
- Graph-Aware Contrasting for Multivariate Time-Series Classification (AAAI 2024 · similarity 0.87)
- Towards Dynamic Spatial-Temporal Graph Learning: A Decoupled Perspective (AAAI 2024 · similarity 0.87)
- Designing Specialized Two-Dimensional Graph Spectral Filters for Spatial-Temporal Graph Modeling (AAAI 2025 · similarity 0.86)
- Alleviating Performance Disparity in Adversarial Spatiotemporal Graph Learning Under Zero-Inflated Distribution (AAAI 2025 · similarity 0.85)
- S3GCL: Spectral, Swift, Spatial Graph Contrastive Learning (ICML 2024 · similarity 0.85)
- Road Network Representation Learning with the Third Law of Geography (NeurIPS 2024 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
