AI Paper Atlas

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

ICML 2026 · Probabilistic and scientific modeling › Molecules and proteins · approach: Machine-learned interatomic potentials

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 changedWhereMeasuredEffect
use 6 multi-shot ED images instead of 1 shot (ablation)VisionED · HOPV15 PCEMAE PCE: 1.596 -> 1.465Improved
increase shots from 6 to 60 (ablation)VisionED · HOPV15 PCEMAE PCE: 1.465 -> 1.450No clear effect
replace 512-point ED point cloud (PointVector) with 6-shot ED imagesVisionED · HOPV15 PCEMAE PCE: 2.031 -> 1.465Improved
increase point cloud size from 512 to 4096 points (ablation)PointVector · HOPV15 PCEMAE PCE: 2.031 -> 2.038No clear effect
pre-train on ED images (VisionED) vs best baseline on acceptor splitOPEP2 acceptor split · OPEP2 acceptor scaffold splitMAE PCE: 2.448 -> 2.271Improved
VisionED vs EquiformerV2 on donor scaffold splitOPEP2 donor split · OPEP2 donor scaffold splitMAE PCE: 3.175 -> 3.034Improved
VisionED vs best baseline on HOPV15 5-fold CVHOPV15 · HOPV15 5-fold CVMAE JSC: 2.750 -> 2.339Improved
VisionED vs VideoMol on solvent scaffold splitDeep4Chem · Deep4Chem solvent scaffold splitMAE λAbs: 32.045 -> 22.502Improved
VisionED vs VideoMol on emitter scaffold splitDeep4Chem · Deep4Chem emitter scaffold splitMAE λAbs: 59.573 -> 51.023Improved
train on OPEP2 and test on HOPV15 (out-of-distribution)VisionED · OPEP2 to HOPV15 OODMAE JSC: 6.8% increase in JSC MAEMixed
train with 50% of data instead of 100% for baselinesVisionED · OPEP2 PCE low-dataMAE PCE: 2.649 -> 2.326Improved
zero-shot ratio-weighted ensemble for ternary blendsVisionED · Ternary OPV blendsPCE accuracy: 92.77%Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy