AI Paper Atlas

Research map / 3D vision / 3D generation

MolAlign3D: Enhancing Fixed-Dimensional E(3)-Equivariant Latent Space for High-Fidelity 3D Molecular Reconstruction and Editing

Zitao Chen, Jiatong Ji, Yinjun Jia, Wei-Ying Ma, Yanyan Lan

ICML 2026 · 3D vision › 3D generation · approach: 2D-to-3D pretrained representations

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
anchor feature latents with pretrained Uni-Mol embeddingsMolFLAE · GEOM-Drugs reconstructionMCS-IoU: 46.11% to 80.79%Improved
anchor feature latents with pretrained Uni-Mol embeddingsMolFLAE · GEOM-Drugs reconstructionShape Tanimoto: 49.84% to 64.91%Improved
anchor feature latents with pretrained Uni-Mol embeddingsMolFLAE · MoleculeNet ClinToxROC-AUC: 75.2 -> 93.7Improved
anchor feature latents with pretrained Uni-Mol embeddingsMolFLAE · MoleculeNet LipophilicityRMSE: 0.835 -> 0.595Improved
anchor feature latents with pretrained Uni-Mol embeddingsMolFLAE · MoleculeNet QM7MAE: 74.3 -> 51.7Improved
anchor feature latents with pretrained Uni-Mol embeddingsMolFLAE · ZINC atom number editing (+0)MCS-IoU: 83.80 -> 87.78Improved
anchor feature latents with pretrained Uni-Mol embeddingsMolFLAE · ZINC atom number editing (-2)MCS-IoU: 70.00 -> 70.26No clear effect
anchor feature latents with pretrained Uni-Mol embeddingsMolFLAE · GEOM-Drugs unconditional generationvalidity: 99.7 -> 99.2No clear effect
anchor feature latents with pretrained Uni-Mol embeddingsMolFLAE · GEOM-Drugs unconditional generationSA: 0.75 -> 0.69Hurt
anchor feature latents with pretrained Uni-Mol embeddingsMolFLAE · Hybrid latent swap, preserving semantic latentMACCS similarity: 58.2 -> 66.8Improved
anchor feature latents with pretrained Uni-Mol embeddingsMolFLAE · Interpolation N=8 sp3fracPearson r: 0.4031 -> 0.5170Improved
use Uni-Mol embeddings alone (no geometry-aware latent) for decodingMolFLAE · GEOM-Drugs reconstructionMCS-IoU: 34.35 vs 80.79Hurt

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