Research map / Graph learning / Graph self-supervised learning
Informative Subgraph Extraction with Deep Reinforcement Learning for Drug-Drug Interaction Prediction
Jiancong Xie, Wentao Wei, Chi Zhang, Jiahua Rao, Yuedong Yang
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 fixed k-hop subgraph extraction with RL-based subgraph sampler (ablation) | RISE-DDI · DrugBank (transductive) | AUC: 0.890 -> 0.926 | Improved |
| replace PageRank-based extractor with RL sampler (ablation) | RISE-DDI · OGB-biokg (transductive) | AUC: 0.947 -> 0.970 | Improved |
| replace RL sampler with fixed sampler (best fixed vs RL) (ablation) | RISE-DDI · KEGG (transductive) | AUC: 0.959 -> 0.971 | Improved |
| add molecular graph features and multi-head attention fusion (structure-aware module) (ablation) | RISE-DDI · three DDI benchmarks, transductive and inductive | AUC | Improved |
| add structure-aware module, inductive vs transductive (ablation) | RISE-DDI · inductive setting | AUC | Improved |
| increase subgraph layers L beyond 2 (ablation) | RISE-DDI · three DDI benchmarks | AUC | Hurt |
| increase RL steps T (subgraph size) (ablation) | RISE-DDI · three DDI benchmarks | AUC | Improved |
| use RL subgraph sampling instead of adaptive subgraph search (CSSE-DDI) in inductive setting | RISE-DDI · OGB-biokg (inductive, both drugs unseen) | AUROC: 0.562 -> 0.705 | Improved |
| use RL subgraph sampler at T=40 versus no sampling | RISE-DDI · inference cost, batch size 128 | inference time: +1.5 seconds per batch | Hurt |
| use RL sampler instead of fixed sampler for large-degree graph (DrugBank) | RISE-DDI · DrugBank (transductive) | AUC: 0.867 -> 0.926 | Improved |
Closest papers by meaning
- Customized Subgraph Selection and Encoding for Drug-drug Interaction Prediction (NeurIPS 2024 · similarity 0.92)
- Dual-Channel Learning Framework for Drug-Drug Interaction Prediction via Relation-Aware Heterogeneous Graph Transformer (AAAI 2024 · similarity 0.90)
- MOTIVE: A Drug-Target Interaction Graph For Inductive Link Prediction (NeurIPS 2024 · similarity 0.88)
- ExDDI: Explaining Drug-Drug Interaction Predictions with Natural Language (AAAI 2025 · similarity 0.87)
- Knowledge Enhanced Representation Learning for Drug Discovery (AAAI 2024 · similarity 0.87)
- Motif-Oriented Representation Learning with Topology Refinement for Drug-Drug Interaction Prediction (AAAI 2025 · 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.
