Research map / Vision-language and multimodal / Cross-modal learning
Generalizable Drug–Target Interaction Prediction via ESM-2 Representations and Progressive Contrastive Curriculum Learning
Qianyang Wu, Jingwei Lv, Zilong Zhang, Feifei Cui
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 ESM-2 protein embeddings instead of handcrafted protein features (ablation) | ESP-DTI · BindingDB (cross-domain) | AUROC: 0.584 -> 0.701 | Improved |
| add ESM-2 protein embeddings instead of handcrafted protein features (ablation) | ESP-DTI · BioSNAP (cross-domain) | AUROC: 0.602 -> 0.763 | Improved |
| add CLIP-style cross-modal alignment loss (ablation) | ESP-DTI · BindingDB (cross-domain) | AUROC: 0.640 -> 0.701 | Improved |
| add progressive confidence-based sample selection curriculum (ablation) | ESP-DTI · BindingDB (cross-domain) | AUROC: 0.639 -> 0.701 | Improved |
| add contrastive refinement on easy/hard partitions (ablation) | ESP-DTI · BindingDB (cross-domain) | AUROC: 0.637 -> 0.701 | Improved |
| add curriculum, contrastive refinement and CLIP alignment on BioSNAP (ablation) | ESP-DTI · BioSNAP (cross-domain) | AUROC: 0.749 -> 0.763 (curriculum) | Improved |
| ESP-DTI versus strongest baseline under cross-domain split | ESP-DTI · BindingDB and BioSNAP (cross-domain) | AUROC: 0.657 -> 0.701 (BindingDB); 0.728 -> 0.763 (BioSNAP) | Improved |
| ESP-DTI versus MlanDTI under BindingDB cold-start | ESP-DTI · BindingDB (cold-start) | AUROC: 0.671 -> 0.666 | No clear effect |
| ESP-DTI versus MlanDTI on Human random split F1 | ESP-DTI · Human (random split) | F1: 0.961 -> 0.944 | Hurt |
| ESP-DTI versus MlanDTI on BioSNAP random split AUROC | ESP-DTI · BioSNAP (random split) | AUROC: 0.909 -> 0.943 | Improved |
Closest papers by meaning
- R-DTI: Drug Target Interaction Prediction Based on Second-Order Relevance Exploration (AAAI 2025 · similarity 0.88)
- Knowledge Enhanced Representation Learning for Drug Discovery (AAAI 2024 · similarity 0.88)
- Blend the Separated: Mixture of Synergistic Experts for Data-Scarcity Drug-Target Interaction Prediction (AAAI 2025 · similarity 0.87)
- M²N: A Progressive Macro-to-Micro 3D Modeling Scheme for Unveiling Drug-Target Affinity (AAAI 2025 · similarity 0.87)
- Multilevel Attention Network with Semi-supervised Domain Adaptation for Drug-Target Prediction (AAAI 2024 · similarity 0.86)
- Omni-Mol: Multitask Molecular Model for Any-to-any Modalities (NeurIPS 2025 · similarity 0.86)
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.
