Research map / Graph learning / GNN architectures
Alternately Optimized Graph Neural Networks
Haoyu Han, Xiaorui Liu, Haitao Mao, MohamadAli Torkamani, Feng Shi, Victor Lee, Jiliang Tang
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 |
|---|---|---|---|
| Alternating optimization with lazy propagation (update F only 5 times) instead of per-epoch propagation (ablation) | ALT-OPT on ogbn-arxiv · ogbn-arxiv | accuracy (%): 72.70 vs 72.76 (Full) | No clear effect |
| Alternating optimization with lazy propagation (update F only 5 times) instead of per-epoch propagation (ablation) | ALT-OPT on ogbn-products · ogbn-products | accuracy (%): 82.64 vs 81.83 (Full) | Improved |
| Reduce F updates to only 1 time during training (ablation) | ALT-OPT on ogbn-arxiv · ogbn-arxiv | accuracy (%): 70.09 vs 72.70 (ALT-OPT-5) | Hurt |
| Increase F-update frequency from 1 to 5 times (ablation) | ALT-OPT on ogbn-products · ogbn-products | accuracy (%): 80.03 (1x) to 82.64 (5x) | Improved |
| Alternating optimization (ALT-OPT-5) vs APPNP end-to-end training | ogbn-products · ogbn-products | training time (s): 222 s vs 1,913 s | Improved |
| Alternating optimization with lazy propagation vs APPNP | ogbn-products · ogbn-products | memory (GB): 24.49 vs 29.15 | Improved |
| Feature diffusion (LP-enhanced features) before MLP pretraining removed (ablation) | ALT-OPT on Cora and CiteSeer · Cora, CiteSeer low label rates | accuracy (%): gap large at low label rates, small at higher rates | Improved |
| Remove pseudo labels (train MLP on labeled data only) (ablation) | ALT-OPT on Cora and CiteSeer · Cora, CiteSeer | accuracy (%): large impact especially at low label rate (figure only) | Improved |
| Replace MSE loss with cross-entropy (ALT-OPT-N with softmax normalization and reweighting) (ablation) | ALT-OPT · PubMed, CS, ogbn-arxiv | accuracy (%): 78.66 vs 79.16 (PubMed); 92.69 vs 92.77 (CS); 72.58 vs 72.60 (arxiv) | No clear effect |
| Use Laplacian filter L^2 (ALT-OPT-H) instead of L on heterophily graphs | ALT-OPT on Chameleon · Chameleon | accuracy (%): 70.62 vs LINKX 68.42 | Improved |
| Use Laplacian filter L^2 (ALT-OPT-H) instead of L on heterophily graphs | ALT-OPT on Squirrel · Squirrel | accuracy (%): 61.56 vs LINKX 61.81 | No clear effect |
| Alternating optimization with pseudo-labels vs end-to-end GNNs at low label rates | ALT-OPT on Cora, CiteSeer, Pubmed · label rate 5 (transductive) | accuracy (%): Cora 76.78 vs APPNP 75.86; CiteSeer 67.48 vs 63.92; Pubmed 73.51 vs 72.68 | Improved |
Closest papers by meaning
- Graph Random Neural Networks for Semi-Supervised Learning on Graphs (NeurIPS 2020 · similarity 0.90)
- Beyond Homophily in Graph Neural Networks: Current Limitations and Effective Designs (NeurIPS 2020 · similarity 0.90)
- Normalize Then Propagate: Efficient Homophilous Regularization for Few-Shot Semi-Supervised Node Classification (AAAI 2025 · similarity 0.89)
- $p$-Laplacian Based Graph Neural Networks (ICML 2022 · similarity 0.88)
- Optimization of Graph Neural Networks: Implicit Acceleration by Skip Connections and More Depth (ICML 2021 · similarity 0.88)
- Fast Graph Sharpness-Aware Minimization for Enhancing and Accelerating Few-Shot Node Classification (NeurIPS 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.
