Research map / Graph learning / Graph algorithms and clustering
Ensemble Quadratic Assignment Network for Graph Matching
Haoru Tan, Chuang Wang, Sitong Wu, Xu-Yao Zhang, Fei Yin, Cheng‐Lin 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 |
|---|---|---|---|
| add channel-mixing ensemble (EQAN) instead of naive averaging (ablation) | DPGM solvers · Synthetic random graphs (ablation, n_in=35, n_out=15, sigma=0.3) | accuracy: 37.0 -> 99.1 | Improved |
| add ensemble of solvers instead of single solver (ablation) | DPGM · Synthetic random graphs (ablation, n_in=35, n_out=15, sigma=0.3) | accuracy: 37.4 -> 99.1 | Improved |
| learn internal solver parameters instead of random (ablation) | EQAN · Synthetic random graphs (ablation, n_in=35, n_out=15, sigma=0.3) | accuracy: 74.1 -> 99.1 | Improved |
| ensemble GAGM solvers instead of single (ablation) | GAGM · Synthetic random graphs (ablation, n_in=35, n_out=15, sigma=0.3) | accuracy: 39.9 -> 96.2 | Improved |
| ensemble SM solvers instead of single (ablation) | SM · Synthetic random graphs (ablation, n_in=35, n_out=15, sigma=0.3) | accuracy: 38.6 -> 94.7 | Improved |
| ensemble RRWM solvers instead of single (ablation) | RRWM · Synthetic random graphs (ablation, n_in=35, n_out=15, sigma=0.3) | accuracy: 40.1 -> 99.4 | Improved |
| use RRWM instead of DPGM as base solver (ablation) | EQAN · Synthetic random graphs (ablation, n_in=35, n_out=15, sigma=0.3) | accuracy: 99.1 -> 99.4 | No clear effect |
| use decision features from all blocks instead of last block only (ablation) | EQAN · Synthetic random graphs (ablation, n_in=35, n_out=15, sigma=0.3) | accuracy: 93.4 -> 99.1 | Improved |
| add Sinkhorn normalization after each layer instead of single Sinkhorn (ablation) | EQAN · Synthetic random graphs (ablation, n_in=35, n_out=15, sigma=0.3) | accuracy: 71.4 -> 99.1 | Improved |
| increase width C from 8 to 32 (ablation) | EQAN · Synthetic random graphs (ablation, n_in=35, n_out=15, sigma=0.3) | accuracy: 79.1 -> 99.1 | Improved |
| increase width C from 32 to 64 (ablation) | EQAN · Synthetic random graphs (ablation, n_in=35, n_out=15, sigma=0.3) | accuracy: 99.1 -> 99.3 | No clear effect |
| increase depth L from 2 to 5 (ablation) | EQAN · Synthetic random graphs (ablation, n_in=35, n_out=15, sigma=0.3) | accuracy: 79.2 -> 99.1 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- SIGMA — SIGMA: Semantic-Complete Graph Matching for Domain Adaptive Object Detection (CVPR 2022) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Stochastic Iterative Graph Matching (ICML 2021 · similarity 0.90)
- Neural Graph Matching Network: Learning Lawler’s Quadratic Assignment Problem with Extension to Hypergraph and Multiple-graph Matching (TPAMI 2021 · similarity 0.90)
- Combinatorial Learning of Robust Deep Graph Matching: An Embedding Based Approach (TPAMI 2020 · similarity 0.89)
- ClusterGNN: Cluster-Based Coarse-To-Fine Graph Neural Network for Efficient Feature Matching (CVPR 2022 · similarity 0.89)
- Learning Partial Graph Matching via Optimal Partial Transport (ICLR 2025 · similarity 0.86)
- Adapting Precomputed Features for Efficient Graph Condensation (ICML 2025 · similarity 0.86)
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.
