Research map / Graph learning / Graph self-supervised learning
GraphTER: Unsupervised Learning of Graph Transformation Equivariant Representations via Auto-Encoding Node-Wise Transformations
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| learn representations by decoding node-wise transformations rather than reconstructing input data | GraphTER · ModelNet40 | accuracy (%): 90.6 -> 92.0 | Helped |
| learn representations by decoding node-wise transformations rather than reconstructing input data | GraphTER · ShapeNet part | mIoU (%): 68.0 -> 81.9; 13.9% | Helped |
| use shearing instead of translation or rotation (ablation) | GraphTER · ModelNet40 | accuracy (%): 89.94; 90.47; 90.99 | Helped |
| use shearing instead of translation or rotation (ablation) | GraphTER · ShapeNet part | mIoU (%): 79.90; 80.35; 81.13 | Helped |
| sample nodes globally rather than locally (ablation) | GraphTER · ModelNet40 | accuracy (%): 90.70; 90.24; 0.46% | Helped |
| sample nodes locally rather than globally (ablation) | GraphTER · ShapeNet part | mIoU (%): 80.53; 80.39; 0.14% | Helped |
| use isotropic rather than anisotropic transformations (ablation) | GraphTER · ModelNet40 | accuracy (%): 91.15; 90.24; 90.68; 89.80 | Helped |
| use isotropic rather than anisotropic transformations (ablation) | GraphTER · ShapeNet part | mIoU (%): 80.64; 80.15; 80.94; 80.12 | Mixed |
| transform a smaller sampled subset of nodes rather than all nodes (ablation) | GraphTER with translation · ModelNet40 | accuracy (%): 89.94; 89.77; 90.46; 89.83 | No clear effect |
Closest papers by meaning
- Structure-Aware Transformer for Graph Representation Learning (ICML 2022 · similarity 0.89)
- VCR-Graphormer: A Mini-batch Graph Transformer via Virtual Connections (ICLR 2024 · similarity 0.88)
- Graph As Point Set (ICML 2024 · similarity 0.87)
- Subgraph Invariant Learning Towards Large-Scale Graph Node Classification (AAAI 2025 · similarity 0.87)
- Translating Subgraphs to Nodes Makes Simple GNNs Strong and Efficient for Subgraph Representation Learning (ICML 2024 · similarity 0.87)
- Transitivity-Preserving Graph Representation Learning for Bridging Local Connectivity and Role-Based Similarity (AAAI 2024 · similarity 0.87)
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.