Research map / Graph learning / Knowledge and temporal graphs
Improving Graph Matching with Positional Reconstruction Encoder-Decoder Network
Yixiao Zhou, Ruiqi Jia, Hongxiang Lin, Hefeng Quan, Yumeng Zhao, Xiaoqing Lyu
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 |
|---|---|---|---|
| add PR-EnDec positional encoding | BBGM · PascalVOC | matching accuracy (%): 80.1 -> 84.1 | Improved |
| add PR-EnDec positional encoding | BBGM · Willow (PT+WT) | matching accuracy (%): 97.4 -> 99.3 | Improved |
| add PR-EnDec positional encoding | BBGM · SPair-71k | matching accuracy (%): 78.9 -> 82.4 | Improved |
| add PR-EnDec positional encoding | BBGM · IMC-PT-SparseGM | matching accuracy (%): 88.4 -> 90.7 | Improved |
| remove positional encoder (ablation) | PREGM · PascalVOC | matching accuracy (%): 84.1% -> 83.4% | Hurt |
| remove spatial relation decoder (ablation) | PREGM · PascalVOC | matching accuracy (%): 84.1% -> 81.7% | Hurt |
| remove area reconstruction (ablation) | PREGM · PascalVOC | matching accuracy (%): 84.1% -> 83.8% | Hurt |
| remove relative distance reconstruction (ablation) | PREGM · PascalVOC | matching accuracy (%): 84.1% -> 82.7% | Hurt |
| remove visual features (spatial only) (ablation) | PREGM · PascalVOC | matching accuracy (%): 84.1% -> 76.8% | Hurt |
| learning rate 6e-4 instead of 9e-4 (ablation) | PREGM · PascalVOC | matching accuracy (%): 84.1% -> 83.2% | Hurt |
| learning rate 10e-4 instead of 9e-4 (ablation) | PREGM · PascalVOC | matching accuracy (%): 84.1% -> 83.9% | No clear effect |
| balance factor lambda 1/128 instead of 1/32 (ablation) | PREGM · PascalVOC | matching accuracy (%): 84.1% -> 83.7% | Hurt |
Closest papers by meaning
- Graph Positional and Structural Encoder (ICML 2024 · similarity 0.86)
- GraphCroc: Cross-Correlation Autoencoder for Graph Structural Reconstruction (NeurIPS 2024 · similarity 0.86)
- Iteratively Refined Early Interaction Alignment for Subgraph Matching based Graph Retrieval (NeurIPS 2024 · similarity 0.85)
- GPEN: Global Position Encoding Network for Enhanced Subgraph Representation Learning (ICML 2025 · similarity 0.85)
- Graph As Point Set (ICML 2024 · similarity 0.85)
- Learnable Spatial-Temporal Positional Encoding for Link Prediction (ICML 2025 · 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.
