AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

Align Your Structures: Generating Trajectories with Structure Pretraining for Molecular Dynamics

Aniketh Iyengar, Jiaqi Han, Pengwei Sun, Mingjian Jiang, Jianwen Xie, Stefano Ermon

ICLR 2026 · Image generation › Autoregressive and latent generation · approach: Equivariant diffusion for 3D molecules · 0 citations (OpenAlex)

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 conformer structure pretraining (ablation)EGInterpolator · QM9 unconditional generationBond length JSD (mean): 0.583 -> 0.210Improved
add conformer structure pretraining (ablation)EGInterpolator · Drugs forward simulationBond length JSD (mean): 0.386 -> 0.142Improved
add conformer structure pretraining (ablation)EGInterpolator · QM9 unconditional generationTICA_0,1 JSD (mean): 0.680 -> 0.636Improved
use cascaded interpolation architecture instead of a residual temporal stack (ablation)EGInterpolator · Drugs forward simulationTorsion JSD (mean): 0.414 -> 0.377Improved
use cascaded interpolation architecture instead of a residual temporal stack (ablation)EGInterpolator · Drugs forward simulationBond length JSD (mean): 0.330 -> 0.142Improved
use Casc variant instead of SimpleEGInterpolator · QM9 unconditional generationBond angle JSD (mean): 0.357 -> 0.305Improved
use Casc variant instead of SimpleEGInterpolator · Drugs forward simulationBond length JSD (mean): 0.258 -> 0.142Improved
condition the first block on a ground-truth frame (forward simulation vs unconditional) (ablation)EGInterpolator-Casc · DrugsBond length JSD (mean): 0.195 -> 0.142Improved
train unconditionally without first frame, compared with GeoTDM forwardEGInterpolator-Casc-U · DrugsBond length JSD (mean): 0.643 -> 0.195Improved
joint training on QM9 and DrugsEGInterpolator-Both · QM9 unconditional generationBond length JSD (mean): 0.210 -> 0.168Improved
select checkpoint 539 over 99 (precision-oriented)BasicES · QM9 conformersCOV-P mean (%): 55.23 -> 58.12Improved
select checkpoint 539 over 99 (precision-oriented)BasicES · QM9 conformersCOV-R mean (%): 90.18 -> 87.62Hurt

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