AI Paper Atlas

Research map / Probabilistic and scientific modeling / Molecules and proteins

Hierarchical Multi-Scale Molecular Conformer Generation with Structural Awareness

Jiapeng Hu, Weizhi Gao, Zhichao Hou, Xiaorui Liu

ICLR 2026 · Probabilistic and scientific modeling › Molecules and proteins · approach: Structure-based molecule design with diffusion

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
add ground-truth heavy-atom geometric guidanceGeoDiff · GEOM-Drugs preliminaryCOV-R mean: 64.12 -> 99.58Improved
add ground-truth heavy-atom geometric guidanceGeoDiff · GEOM-Drugs preliminaryMAT-R mean: 1.1444 -> 0.5035Improved
add MSGEN two-stage generationGeoDiff · GEOM-DrugsCOV-P mean: 60.17 -> 66.26Improved
add MSGEN two-stage generationGeoDiff · GEOM-DrugsMAT-R mean: 0.8686 -> 0.8424Improved
add MSGENGeoDiff · GEOM-QM9 ensemble energyMAE average energy (eV): 0.2597 -> 0.1795Improved
add MSGEN under domain transfer (Drugs to QM9)GeoDiff · GEOM-QM9COV-R mean: 74.94 -> 83.73Improved
add third stage (Murcko scaffold) to MSGEN (ablation)ET-Flow · GEOM-DrugsCOV-R mean: 80.50 -> 81.91Improved
add third stage (Murcko scaffold) to MSGEN (ablation)GeoDiff · GEOM-DrugsCOV-R mean: 90.41 -> 91.05Improved
add MSGEN 2-stage to ConfGFConfGF · GEOM-DrugsCOV-R mean: 62.15 -> 65.51Improved
add MSGEN 2-stage to EBDEBD · GEOM-DrugsCOV-R mean: 92.10 -> 91.92No clear effect
add second stage (hierarchical design) after Stage 1 (ablation)GeoDiff+MSGEN · GEOM-Drugs coarse atomsCOV-R: 87.86 -> 89.17 -> 90.41Improved
use molecular upsampling instead of random or centroid sampling (ablation)MSGEN · GEOM-DrugsCOV/MATImproved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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