Research map / Graph learning / Graph algorithms and clustering
Stochastic Iterative Graph Matching
Linfeng Liu, Michael C Hughes, Soha Hassoun, Liping Liu
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 |
|---|---|---|---|
| Use SIGMA instead of S-GWL | noisy BA and PPI graph matching · BA(500), BA(1000), BA(2000), and PPI with 5% noise | node correctness (%): 32.0±1.5 -> 93.8±0.3 (BA(500)); 22.9±0.3 -> 97.3±0.2 (BA(1000)); 16.2±0.4 -> 99.0±0.1 (BA(2000)); 81.1 -> 84.7±0.4 (PPI) | Improved |
| Add iterative refinement (T=4) rather than T=0 (ablation) | SIGMA · PPI, 5% noise | node correctness (%): 83.1±0.3 -> 84.7±0.4 | Improved |
| Add iterative refinement (T=4) rather than T=0 (ablation) | SIGMA · BA(500), BA(1000), and BA(2000), 5% noise | node correctness (%): 93.5±0.1 -> 93.8±0.3; 97.2±0.1 -> 97.3±0.2; 98.9±0.1 -> 99.0±0.1 | No clear effect |
| Add the stochastic framework (ablation) | SIGMA · Noisy BA and PPI graph matching | node correctness (%): 93.6±0.1 -> 93.8±0.3 (BA(500)); 97.1±0.2 -> 97.3±0.2 (BA(1000)); 98.9±0.1 -> 99.0±0.1 (BA(2000)); 84.1±0.4 -> 84.7±0.4 ( | Mixed |
| Add dummy nodes (ablation) | SIGMA · Noisy BA and PPI graph matching | node correctness (%): 50.3±4.1 -> 93.8±0.3 (BA(500)); 75.0±5.9 -> 97.3±0.2 (BA(1000)); 80.4±3.1 -> 99.0±0.1 (BA(2000)); 83.2±0.2 -> 84.7±0.4 ( | Improved |
| Use SIGMA instead of DGMC | RDM pattern matching · RDM pattern matching | node correctness (%): 66.9±0.9 -> 78.3±0.8 | Improved |
| Add iterative refinement rather than T=0 (ablation) | SIGMA · RDM pattern matching | node correctness (%): 76.4±0.8 -> 78.3±0.8 | Improved |
| Add the stochastic framework (ablation) | SIGMA · RDM pattern matching | node correctness (%): 76.0±0.8 -> 78.3±0.8 | Improved |
| Add dummy nodes (ablation) | SIGMA · RDM pattern matching | node correctness (%): 63.6±0.9 -> 78.3±0.8 | Improved |
| Use the supervised objective in addition to the QAP objective (ablation) | SIGMA for RDM pattern matching · RDM pattern matching | node correctness (%): 66.2±1.0 -> 78.3±0.8 | Improved |
| Use SIGMA instead of DGMC | PASCAL VOC keypoint matching · PASCAL VOC, Berkeley annotations | Hit@1 (%): 73.0 -> 76.2 | Improved |
| Add stochastic matching rather than optimizing a single matching (ablation) | SIGMA · PASCAL VOC keypoint matching | Hit@1 (%): 73.1 -> 76.2 | Improved |
Closest papers by meaning
- Ensemble Quadratic Assignment Network for Graph Matching (IJCV 2024 · similarity 0.90)
- Combinatorial Learning of Robust Deep Graph Matching: An Embedding Based Approach (TPAMI 2020 · similarity 0.90)
- Deep Learning of Graph Matching (CVPR 2018 · similarity 0.89)
- Learning Partial Graph Matching via Optimal Partial Transport (ICLR 2025 · similarity 0.89)
- Deep Learning of Partial Graph Matching via Differentiable Top-K (CVPR 2023 · similarity 0.89)
- Deep Latent Graph Matching (ICML 2021 · similarity 0.89)
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.
