Research map / Graph learning / Knowledge and temporal graphs
Kernelized Edge Attention: Addressing Semantic Attention Blurring in Temporal Graph Neural Networks
Govind Waghmare, Srini Rohan Gujulla Leel, Nikhil Tumbde, Sumedh B G, Sonia Gupta, Srikanta Bedathur
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 Laplacian kernel modulation of edge features (ablation) | TGN · tgbl-wiki (test) | MRR: 0.396 -> 0.474 | Improved |
| add RBF kernel modulation (ablation) | TGN · tgbl-coin (test) | MRR: 0.586 -> 0.636 | Improved |
| add Laplacian kernel modulation (ablation) | TGN · tgbl-review (test) | MRR: 0.349 -> 0.378 | Improved |
| use learned MLP kernel instead of Laplacian (ablation) | KEAT-TGN · tgbl-coin (test) | MRR: 0.632 -> 0.617 | Hurt |
| use RBF instead of Laplacian kernel (ablation) | KEAT-TGN · tgbl-wiki (test) | MRR: 0.474 -> 0.437 | Hurt |
| add KEAT kernel attention | DyGFormer · tgbl-review (test) | MRR: 0.224 -> 0.412 | Improved |
| add KEAT kernel attention | DyGFormer · tgbl-comment (test) | MRR: 0.670 -> 0.776 | Improved |
| add KEAT kernel attention | DyGFormer · tgbl-wiki (test) | MRR: 0.798 -> 0.815 | Improved |
| add KEAT kernel attention | TGN · tgbl-comment (test) | MRR: 0.379 -> 0.400 | Improved |
| add KEAT kernel attention | TGN · tgbn-trade (test) | NDCG@10: +5.30% | Improved |
| add KEAT kernel attention | TGN · tgbn-token (test) | NDCG@10: +1.1% | Improved |
| add KEAT on top of different time encodings (TGN/TGAT, LeTE) | TGN · tgbl-wiki (test) | MRR: +7.8% (TGN/TGAT), +7.3% (LeTE) | Improved |
Closest papers by meaning
- Simple and Efficient Heterogeneous Temporal Graph Neural Network (NeurIPS 2025 · similarity 0.89)
- Unifying Text Semantics and Graph Structures for Temporal Text-attributed Graphs with Large Language Models (NeurIPS 2025 · similarity 0.87)
- Temporal Graph Neural Tangent Kernel with Graphon-Guaranteed (NeurIPS 2024 · similarity 0.87)
- Attention Mechanisms Perspective: Exploring LLM Processing of Graph-Structured Data (ICML 2025 · similarity 0.87)
- Graph Attention is Not Always Beneficial: A Theoretical Analysis of Graph Attention Mechanisms via Contextual Stochastic Block Models (ICML 2025 · similarity 0.87)
- HONGAT: Graph Attention Networks in the Presence of High-Order Neighbors (AAAI 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.
