Research map / Probabilistic and scientific modeling / Molecules and proteins
Physics-Informed Pre-training on Efficient Electron-Density Images for Organic Material Property Prediction
Zhixiang Cheng, Hongxin Xiang, Mingquan Liu, Tengfei Ma, Yingzhuo Tu, Wenjie Du, Bosheng Song, Yiping Liu, xiangxiang Zeng
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 |
|---|---|---|---|
| use 6 multi-shot ED images instead of 1 shot (ablation) | VisionED · HOPV15 PCE | MAE PCE: 1.596 -> 1.465 | Improved |
| increase shots from 6 to 60 (ablation) | VisionED · HOPV15 PCE | MAE PCE: 1.465 -> 1.450 | No clear effect |
| replace 512-point ED point cloud (PointVector) with 6-shot ED images | VisionED · HOPV15 PCE | MAE PCE: 2.031 -> 1.465 | Improved |
| increase point cloud size from 512 to 4096 points (ablation) | PointVector · HOPV15 PCE | MAE PCE: 2.031 -> 2.038 | No clear effect |
| pre-train on ED images (VisionED) vs best baseline on acceptor split | OPEP2 acceptor split · OPEP2 acceptor scaffold split | MAE PCE: 2.448 -> 2.271 | Improved |
| VisionED vs EquiformerV2 on donor scaffold split | OPEP2 donor split · OPEP2 donor scaffold split | MAE PCE: 3.175 -> 3.034 | Improved |
| VisionED vs best baseline on HOPV15 5-fold CV | HOPV15 · HOPV15 5-fold CV | MAE JSC: 2.750 -> 2.339 | Improved |
| VisionED vs VideoMol on solvent scaffold split | Deep4Chem · Deep4Chem solvent scaffold split | MAE λAbs: 32.045 -> 22.502 | Improved |
| VisionED vs VideoMol on emitter scaffold split | Deep4Chem · Deep4Chem emitter scaffold split | MAE λAbs: 59.573 -> 51.023 | Improved |
| train on OPEP2 and test on HOPV15 (out-of-distribution) | VisionED · OPEP2 to HOPV15 OOD | MAE JSC: 6.8% increase in JSC MAE | Mixed |
| train with 50% of data instead of 100% for baselines | VisionED · OPEP2 PCE low-data | MAE PCE: 2.649 -> 2.326 | Improved |
| zero-shot ratio-weighted ensemble for ternary blends | VisionED · Ternary OPV blends | PCE accuracy: 92.77% | Improved |
Closest papers by meaning
- EDBench: Large-Scale Electron Density Data for Molecular Modeling (NeurIPS 2025 · similarity 0.87)
- Learning the Electronic Hamiltonian of Large Atomic Structures (ICML 2025 · similarity 0.86)
- A Denoising Pre-training Framework for Accelerating Novel Material Discovery (AAAI 2025 · similarity 0.86)
- A Recipe for Charge Density Prediction (NeurIPS 2024 · similarity 0.86)
- RingFormer: A Ring-Enhanced Graph Transformer for Organic Solar Cell Property Prediction (AAAI 2025 · similarity 0.86)
- Crystalformer: Infinitely Connected Attention for Periodic Structure Encoding (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.
