Research map / Image generation / Autoregressive and latent generation
Quotient-Space Diffusion Model
Yixian Xu, Yusong Wang, Shengjie Luo, Kaiyuan Gao, Tianyu He, Di He, Chang Liu
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 quotient-space diffusion (horizontal projection plus correction term) | ET-Flow(SO(3)) · GEOM-QM9 | Recall AMR mean (A): 0.076 -> 0.069 | Improved |
| add quotient-space diffusion | ET-Flow(SO(3)) · GEOM-DRUGS | Recall coverage mean (%): 74.91 -> 78.50 | Improved |
| add quotient-space diffusion | ET-Flow(O(3)) · GEOM-DRUGS | Precision coverage mean (%): 66.24 -> 72.70 | Improved |
| add quotient-space diffusion | ET-Flow(O(3)) (published number) · GEOM-DRUGS | Precision coverage mean (%): 74.38 -> 72.70 | Hurt |
| use GeoDiff alignment instead of quotient-space diffusion (ablation) | ET-Flow(SO(3)) · GEOM-DRUGS | Precision coverage mean (%): 67.27 -> 59.58 (vs vanilla) | Hurt |
| use AF3 alignment (ablation) | ET-Flow(SO(3)) · GEOM-QM9 | Precision coverage mean (%): 92.10 -> 84.38 | Hurt |
| use AF3 alignment (ablation) | Proteina M_FS^small · Protein generation, ODE sampling | Designability (%): 13.8 -> 3.8 | Hurt |
| add quotient-space diffusion | Proteina M_FS^small · Protein generation, SDE gamma=0.45 | FPSD vs PDB: 332.9 -> 244.5 | Improved |
| add quotient-space diffusion to 60M model vs larger 200M model | Proteina M_FS^small · Protein generation, ODE sampling | FPSD vs PDB: 85.4 (200M) -> 69.9 (60M + QSD) | Improved |
| add quotient-space diffusion to AF3-like alignment-free sampler | Proteina M_FS^small · Protein generation, ODE sampling | fJSD vs AFDB: 0.12 -> 0.11 | No clear effect |
Closest papers by meaning
- Navigating the Design Space of Equivariant Diffusion-Based Generative Models for De Novo 3D Molecule Generation (ICLR 2024 · similarity 0.89)
- Equivariant Diffusion for Crystal Structure Prediction (ICML 2024 · similarity 0.89)
- Diffusion Generative Modeling on Lie Group Representations (NeurIPS 2025 · similarity 0.89)
- SymDrift: One-Shot Generative Modeling under Symmetries (NeurIPS 2026 · similarity 0.89)
- Equivariant Neural Diffusion for Molecule Generation (NeurIPS 2024 · similarity 0.89)
- Scalable Non-Equivariant 3D Molecule Generation via Rotational Alignment (ICML 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.
