Research map / Label-efficient and robust learning / Test-time adaptation
Test-Time Adaptation without Source Data for Out-of-Domain Bioactivity Prediction
Yiming Yang, Zhiyuan Zhou, Yueming Yin, Hoi-Yeung Li, Adams Wai-Kin Kong
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 |
|---|---|---|---|
| apply TAB test-time adaptation without source data | DTIGN ERM model · DTIGN (avg) | Pearson R: 0.414 -> 0.448 | Improved |
| apply TAB on protein OOD | DTIGN model · SIU 0.6 Ki | Kendall tau: 0.060 -> 0.115 | Improved |
| apply TAB on assay OOD | DTIGN model · DrugOOD assay | RMSE: 1.506 -> 1.552 | Hurt |
| apply TAB on assay OOD | DTIGN model · DrugOOD assay | Pearson R: 0.119 -> 0.388 | Improved |
| apply TAB on Kd task | DTIGN model · SIU 0.6 Kd | MAE: 1.028 -> 1.055 | No clear effect |
| add contrastive learning to consistency learning (ablation) | TAB · DTIGN (avg) | Pearson R: 0.432 -> 0.448 | Improved |
| add consistency learning to contrastive learning (ablation) | TAB · DTIGN (avg) | RMSE: 1.201 -> 1.157 | Improved |
| use either component alone instead of ERM (ablation) | TAB variants · DTIGN (avg) | Kendall tau: 0.295 -> 0.285 | No clear effect |
| add uncertainty weighting to consistency loss (ablation) | TAB · DTIGN I2 | RMSE: 1.609 -> 1.481 | Improved |
| add EMA momentum update (ablation) | TAB · DTIGN I1 | Pearson R: 0.572 -> 0.614 | Improved |
| use atom masking instead of coordinate jitter (ablation) | TAB · DTIGN I1 | Pearson R: 0.503 -> 0.614 | Improved |
| use atom masking instead of atom substitution (ablation) | TAB · DTIGN I1 | Pearson R: 0.592 -> 0.614 | Improved |
Closest papers by meaning
- Monroe: A Molecular Foundation Model for In-context Probabilistic Inference (NeurIPS 2026 · similarity 0.87)
- Towards precision protein-ligand affinity prediction benchmark: A Complete and Modification-Aware DAVIS Dataset (NeurIPS 2025 · similarity 0.86)
- Enhancing Ligand Validity and Affinity in Structure-Based Drug Design with Multi-Reward Optimization (ICML 2025 · similarity 0.86)
- Removing Biases from Molecular Representations via Information Maximization (ICLR 2024 · similarity 0.85)
- Drug-TTA: Test-Time Adaptation for Drug Virtual Screening via Multi-task Meta-Auxiliary Learning (ICML 2025 · similarity 0.85)
- MF-LAL: Drug Compound Generation Using Multi-Fidelity Latent Space Active Learning (ICML 2025 · 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.
