Research map / Graph learning / GNN architectures
AutoGEL: An Automated Graph Neural Network with Explicit Link Information
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 inter-layer search (layer connectivity and aggregation) (ablation) | AutoGEL · LP on PB | AUC: 97.16$\pm$0.13 -> 97.27$\pm$0.15 | Helped |
| add inter-layer search space (ablation) | AutoGEL · PROTEINS | accuracy (%): 77.96 -> 82.68 | Helped |
| searched pooling (sum/mean/max) instead of fixed difference pooling (ablation) | AutoGEL · LP on PB | AUC: 96.12$\pm$0.21 -> 97.27$\pm$0.15 | Helped |
| searched pooling instead of fixed difference pooling (ablation) | AutoGEL · LP on C.ele | AUC: 90.38$\pm$0.64 -> 92.90$\pm$1.02 | Helped |
| neighbor-type specific weight matrices instead of one shared matrix (ablation) | AutoGEL · LP on homogeneous graphs | AUC: 97.23$\pm$0.07 -> 97.27$\pm$0.15 (PB) | Helped |
| direction-aware edge filters instead of single matrix (ablation) | AutoGEL · WN18RR | MRR: .470 -> .479 | Helped |
| add edge embedding h_e and composition operator (ablation) | AutoGEL · WN18RR | MRR: .454 -> .479 | Helped |
| add edge embedding h_e (ablation) | AutoGEL · FB15k-237 | MRR: .355 -> .357 | Helped |
| SNAS stochastic search instead of DARTS (ablation) | AutoGEL · LP on PB | AUC: 95.84$\pm$0.29 -> 97.27$\pm$0.15 | Helped |
| SNAS stochastic search instead of DARTS (ablation) | AutoGEL · IMDB-M | accuracy (%): 47.42 -> 56.80 | Helped |
| SNAS instead of DARTS (ablation) | AutoGEL · Cora/CiteSeer/PubMed search time | search time (s): 97 -> 19 (PubMed) | Helped |
| use AutoGEL instead of SANE on node classification | AutoGNN · PubMed | accuracy (%): 90.47 -> 89.68 | Mixed |
Closest papers by meaning
- Autobahn: Automorphism-based Graph Neural Nets (NeurIPS 2021 · similarity 0.88)
- AutoSGNN: Automatic Propagation Mechanism Discovery for Spectral Graph Neural Networks (AAAI 2025 · similarity 0.88)
- AutoTransfer: AutoML with Knowledge Transfer - An Application to Graph Neural Networks (ICLR 2023 · similarity 0.88)
- Graph Neural Networks for Link Prediction with Subgraph Sketching (ICLR 2023 · similarity 0.87)
- A Topological Perspective on Demystifying GNN-Based Link Prediction Performance (ICLR 2024 · similarity 0.85)
- NetInfoF Framework: Measuring and Exploiting Network Usable Information (ICLR 2024 · similarity 0.85)
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.