Research map / Image generation / Fast diffusion sampling
Functional Adjoint Sampler: Scalable Sampling on Infinite Dimensional Spaces
Byoungwoo Park, Juho Lee, Guan-Horng 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 |
|---|---|---|---|
| use FAS function-space sampler instead of TR-LV | transition path sampling · Alanine dipeptide | THP (%): 61.25 -> 100.00 | Improved |
| use FAS instead of TR-LV | transition path sampling · Alanine dipeptide | ETS (kJ/mol): 49.11 -> -29.46 | Improved |
| use FAS instead of TR-LV | transition path sampling · Chignolin | ETS (kJ/mol): -303.98 -> -360.85 | Improved |
| use FAS instead of TR-LV | transition path sampling · Chignolin | THP (%): 43.95 -> 100.00 | Improved |
| use FAS instead of SMD | transition path sampling · Chignolin | RMSD (A): 0.85 -> 0.00 | Improved |
| refine test-time discretization by 100x without retraining (ablation) | FAS · Muller-Brown | ETS (kJ/mol): -36.70 -> -37.83 | No clear effect |
| refine test-time discretization by 100x (ablation) | FAS · Muller-Brown | LLK: 11.48 -> 11.72 | No clear effect |
| refine test-time discretization by 4x (ablation) | FAS · Alanine dipeptide | ETS (kJ/mol): -29.46 -> -33.58 | No clear effect |
| use Feynman-Kac path distribution rather than Langevin-gradient TPS energy as terminal cost (ablation) | FAS · molecular systems | training stability | Improved |
Closest papers by meaning
- Adjoint Sampling: Highly Scalable Diffusion Samplers via Adjoint Matching (ICML 2025 · similarity 0.94)
- Non-equilibrium Annealed Adjoint Sampler (NeurIPS 2025 · similarity 0.91)
- Adjoint Schrödinger Bridge Sampler (NeurIPS 2025 · similarity 0.90)
- On scalable and efficient training of diffusion samplers (NeurIPS 2025 · similarity 0.88)
- AdjointDEIS: Efficient Gradients for Diffusion Models (NeurIPS 2024 · similarity 0.87)
- Transport meets Variational Inference: Controlled Monte Carlo Diffusions (ICLR 2024 · 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.
