Research map / Image generation / Autoregressive and latent generation
Pre-Training Protein Encoder via Siamese Sequence-Structure Diffusion Trajectory Prediction
Zuobai Zhang, Minghao Xu, Aurelie Lozano, Vijil Chenthamarakshan, Payel Das, Jian Tang
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 SiamDiff pre-training | GearNet-Edge atom-level encoder · ATOM3D RES | Accuracy: 0.441$\pm$0.001 -> 0.460$\pm$0.001 | Improved |
| Add SiamDiff pre-training | GearNet-Edge residue-level encoder · ATOM3D PSR | Global ρ: 0.763$\pm$0.012 -> 0.856$\pm$0.007 | Improved |
| Add SiamDiff pre-training | GearNet-Edge residue-level encoder · ATOM3D MSP | AUROC: 0.644$\pm$0.023 -> 0.700$\pm$0.043 | Mixed |
| Add sequence diffusion (ablation) | SiamDiff · ATOM3D PIP atom-level | AUROC: 0.873$\pm$0.004 -> 0.884$\pm$0.003 | Improved |
| Add structure diffusion (ablation) | SiamDiff · ATOM3D MSP atom-level | AUROC: 0.652$\pm$0.021 -> 0.698$\pm$0.020 | Improved |
| Add mutual information maximization between conformers (ablation) | DiffPreT · ATOM3D MSP atom-level | AUROC: 0.680$\pm$0.018 -> 0.698$\pm$0.020 | Improved |
| Use multiscale noise rather than small noise only (ablation) | SiamDiff · ATOM3D PIP atom-level | AUROC: 0.875$\pm$0.002 -> 0.884$\pm$0.003 | Improved |
| Use multiscale noise rather than large noise only (ablation) | SiamDiff · ATOM3D MSP atom-level | AUROC: 0.683$\pm$0.020 -> 0.698$\pm$0.020 | Improved |
| Use small noise only (ablation) | SiamDiff · ATOM3D PSR atom-level | Global ρ: 0.828$\pm$0.005 vs 0.829$\pm$0.012 | No clear effect |
| Add SiamDiff pre-training | ESM-GearNet · ATOM3D PSR residue-level | Mean ρ: 0.595$\pm$0.010 -> 0.656$\pm$0.011 | Improved |
Closest papers by meaning
- Diffusion on Language Model Encodings for Protein Sequence Generation (ICML 2025 · similarity 0.88)
- Steering Masked Discrete Diffusion Models via Discrete Denoising Posterior Prediction (ICLR 2025 · similarity 0.88)
- Diffusion Language Models Are Versatile Protein Learners (ICML 2024 · similarity 0.88)
- 4D Diffusion for Dynamic Protein Structure Prediction with Reference and Motion Guidance (AAAI 2025 · similarity 0.88)
- DPLM-2: A Multimodal Diffusion Protein Language Model (ICLR 2025 · similarity 0.87)
- Simultaneous Modeling of Protein Conformation and Dynamics via Autoregression (NeurIPS 2025 · similarity 0.86)
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.
