Research map / Probabilistic and scientific modeling / Dynamical systems
Beyond Ensembles: Simulating All-Atom Protein Dynamics in a Learned Latent Space
Aditya Sengar, Jiying Zhang, Pierre Vandergheynst, PATRICK BARTH
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 |
|---|---|---|---|
| use autoregressive NN propagator instead of Koopman (ablation) | GLDP latent · ADP backbone dihedral JSD | JSD: 0.138 -> 0.061 | Improved |
| use autoregressive NN instead of Langevin (ablation) | GLDP latent · A1AR backbone dihedral JSD | JSD: 0.052 -> 0.019 | Improved |
| use score-guided Langevin instead of autoregressive NN for side chains (ablation) | GLDP latent · A1AR side-chain chi JSD | JSD: 0.067 -> 0.058 | Improved |
| use autoregressive NN instead of Langevin for rollout stability (ablation) | GLDP latent · ADP rollout failure time | frames: 206 -> 10,000 | Improved |
| use autoregressive NN instead of Koopman for rollout stability (ablation) | GLDP latent · A1AR rollout failure time | frames: 5,740 -> 10,000 | Improved |
| use GLDP-NN instead of LSS | latent simulator · A1AR TICA stability | TICA: Unstable -> 650.20 | Improved |
| use GLDP-NN instead of MD-Gen | latent simulator · A1AR backbone coordinate JSD | JSD: 0.106 -> 0.063 | Improved |
| use GLDP-NN instead of MD-Gen | latent simulator · 7JFL contact correlation | correlation: 0.960 -> 0.949 | Hurt |
| use GLDP-NN instead of LSS | latent simulator · ADP TICA t1 | timescale: 15.80 -> 10.14 | Hurt |
| use GLDP-NN instead of LSS | latent simulator · 7JFL TICA t1 | timescale: 610.50 -> 226.40 | Hurt |
| use GLDP-NN instead of MD-Gen | latent simulator · 1R6W TICA decorrelation | steps: 980 -> 1650 | Improved |
| use Koopman propagator (ablation) | GLDP latent · A2AR activation surface | coverage | Hurt |
Closest papers by meaning
- Generative Modeling of Full-Atom Protein Conformations using Latent Diffusion on Graph Embeddings (NeurIPS 2025 · similarity 0.92)
- TEMPO: Temporal Multi-scale Autoregressive Generation of Protein Conformational Ensembles (NeurIPS 2025 · similarity 0.91)
- Predicting directional flexibility in proteins (NeurIPS 2026 · similarity 0.89)
- Simultaneous Modeling of Protein Conformation and Dynamics via Autoregression (NeurIPS 2025 · similarity 0.88)
- Action-Minimization Meets Generative Modeling: Efficient Transition Path Sampling with the Onsager-Machlup Functional (ICML 2025 · similarity 0.88)
- Exploit Your Latents: Coarse-Grained Protein Backmapping with Latent Diffusion Models (AAAI 2025 · similarity 0.88)
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.
