Research map / Image generation / Autoregressive and latent generation
Scalable Spatio-Temporal SE(3) Diffusion for Long-Horizon Protein Dynamics
Nima Shoghi, Yuxuan Liu, Yuning Shen, Rob Brekelmans, Pan Li, Quanquan Gu
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 |
|---|---|---|---|
| add contextual noise perturbation of history frames (ablation) | STAR-MD · ATLAS 100 ns | combined validity (%): 76.12 -> 85.18 | Improved |
| add contextual noise perturbation of history frames (ablation) | STAR-MD · ATLAS 100 ns | VAMP-2 deviation: 1.22 -> 0.12 | Improved |
| use joint spatio-temporal attention instead of separate space-then-time attention (ablation) | STAR-MD · ATLAS 100 ns | conformation recall: 0.46 -> 0.57 | Improved |
| use joint spatio-temporal attention instead of separate space-then-time attention (ablation) | STAR-MD · ATLAS 100 ns | combined validity (%): 86.70 -> 85.18 | No clear effect |
| place S×T attention inside the diffusion block rather than as external conditioner (ablation) | STAR-MD · ATLAS 100 ns | conformation recall: 0.48 -> 0.57 | Improved |
| use STAR-MD instead of ConfRover | ConfRover · ATLAS 100 ns | combined validity (%): 52.06 -> 85.29 | Improved |
| use STAR-MD instead of MDGen | MDGen · ATLAS 100 ns | conformation recall: 0.28 -> 0.54 | Improved |
| use windowed attention (ConfRover-W) instead of full attention (ablation) | ConfRover · ATLAS 100 ns | combined validity (%): 52.02 -> 45.70 | Hurt |
Closest papers by meaning
- TEMPO: Temporal Multi-scale Autoregressive Generation of Protein Conformational Ensembles (NeurIPS 2025 · similarity 0.90)
- Atomic Trajectory Modeling with State Space Models for Biomolecular Dynamics (NeurIPS 2026 · similarity 0.89)
- Simultaneous Modeling of Protein Conformation and Dynamics via Autoregression (NeurIPS 2025 · similarity 0.89)
- SE(3)-Stochastic Flow Matching for Protein Backbone Generation (ICLR 2024 · similarity 0.89)
- Protein Conformation Generation via Force-Guided SE(3) Diffusion Models (ICML 2024 · similarity 0.88)
- GeoAda: Efficiently Finetune Geometric Diffusion Models with Equivariant Adapters (NeurIPS 2025 · similarity 0.87)
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.
