Research map / Graph learning / Graph self-supervised learning
Towards Open-World Feature Extrapolation: An Inductive Graph Learning Approach
Qitian Wu, Chenxiao Yang, Junchi Yan
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 |
|---|---|---|---|
| replace mean pooling with GNN for new-feature embeddings | FATE-NN · Avazu CTR | ROC-AUC: 0.684 -> 0.710 | Improved |
| replace mean pooling with GNN for new-feature embeddings | FATE-DeepFM · Criteo CTR | ROC-AUC: 0.774 -> 0.783 | Improved |
| add GNN feature extrapolation over base using only observed features | NN · Avazu CTR | ROC-AUC: 0.693 -> 0.710 | Improved |
| add GNN feature extrapolation over base | DeepFM · Criteo CTR | ROC-AUC: 0.773 -> 0.783 | Improved |
| average, KNN or pooling aggregation of observed embeddings for unseen features instead of GNN | NN · UCI datasets | accuracy: FATE 29.8% higher accuracy | Hurt |
| FATE vs incremental retraining on new features (INL-NN) | FATE · UCI datasets | accuracy: +4.1% on average | Improved |
| add DropEdge regularization (ablation) | FATE · UCI datasets | accuracy | Improved |
| asynchronous updates of backbone and GNN instead of joint training (ablation) | FATE · UCI datasets | accuracy | Improved |
| n-fold split vs k-shot sampling training (ablation) | FATE · UCI datasets | accuracy | Mixed |
Closest papers by meaning
- How Neural Networks Extrapolate: From Feedforward to Graph Neural Networks (ICLR 2021 · similarity 0.86)
- Size-Invariant Graph Representations for Graph Classification Extrapolations (ICML 2021 · similarity 0.85)
- Fully-inductive Node Classification on Arbitrary Graphs (ICLR 2025 · similarity 0.85)
- Graph Structure Extrapolation for Out-of-Distribution Generalization (ICML 2024 · similarity 0.84)
- What Improves the Generalization of Graph Transformers? A Theoretical Dive into the Self-attention and Positional Encoding (ICML 2024 · similarity 0.84)
- NLGT: Neighborhood-based and Label-enhanced Graph Transformer Framework for Node Classification (AAAI 2025 · similarity 0.84)
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.
