Research map / Graph learning / Graph self-supervised learning
From Static to Active: Knowledge-Aware Node State Selection in Multi-view Graph Learning
Weiran Liao, Jielong Lu, Yuhong Chen, Shide Du, Hongrong Chen, Shiping Wang
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 activation sequences (mask-based dynamic attention) instead of static kNN graph (ablation) | KAMSSM · Flickr | ACC: 64.5 -> 73.4 | Improved |
| add activation sequences instead of static kNN graph (ablation) | KAMSSM · BDGP | ACC: 95.6 -> 98.6 | Improved |
| add static-sequence history pool and synaptic signals (ablation) | KAMSSM · WebKB | ACC: 78.1 -> 84.6 | Improved |
| add static-sequence history pool and synaptic signals (ablation) | KAMSSM · BDGP | ACC: 97.9 -> 98.6 | Improved |
| add directional inter-view diffusion instead of attention-weighted fusion (ablation) | KAMSSM · OutScene | ACC: 77.5 -> 80.1 | Improved |
| add directional inter-view diffusion instead of attention-weighted fusion (ablation) | KAMSSM · WebKB | ACC: 82.0 -> 84.6 | Improved |
| add directional inter-view diffusion instead of attention-weighted fusion (ablation) | KAMSSM · BDGP | ACC: 97.7 -> 98.6 | Improved |
| KAMSSM versus DSRL on Caltech | KAMSSM · Caltech (10% labels) | ACC: 95.2 (DSRL) vs 95.1 | No clear effect |
| use more views (four vs triple) in multi-view training | graph multi-view baseline on HW · HW | accuracy: 86.2% (dual) -> 83.1% (triple) -> 89.2% (four) | Mixed |
| increase kNN neighbors from 5 to 10 in a single-view graph | base kNN graph · HW | accuracy: 76.5% -> 75.3% | Hurt |
Closest papers by meaning
- Multi-View Graph Clustering via Node-Guided Contrastive Encoding (ICML 2025 · similarity 0.85)
- Towards Continuous Reuse of Graph Models via Holistic Memory Diversification (ICLR 2025 · similarity 0.84)
- EASEMVC:Efficient Dual Selection Mechanism for Deep Multi-View Clustering (CVPR 2025 · similarity 0.84)
- A Topology-aware Graph Coarsening Framework for Continual Graph Learning (NeurIPS 2024 · similarity 0.84)
- Multi-View Incremental Learning with Structured Hebbian Plasticity for Enhanced Fusion Efficiency (AAAI 2025 · similarity 0.84)
- A Dynamical Systems-Inspired Pruning Strategy for Addressing Oversmoothing in Graph Attention Networks (ICML 2025 · similarity 0.83)
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.
