Enhanced Diffusion Sampling: Efficient Rare Event Sampling and Free Energy Calculation with Diffusion Models
Yu Xie, Ludwig Winkler, Lixin Sun, Sarah Lewis, Adam E. FosterJosé Jiménez Luna, Tim Hempel, Michael Gastegger, Yaoyi Chen,Iryna Zaporozhets, Cecilia Clementi, Christopher M. Bishop, Frank Noé Microsoft Research AI for Science · 2026
Read Enhanced Diffusion Sampling: Efficient Rare Event Sampling and Free Energy Calculation with Diffusion Models by Yu Xie, Ludwig Winkler, Lixin Sun, Sarah Lewis, Adam E. FosterJosé Jiménez Luna, Tim Hempel, Michael Gastegger, Yaoyi Chen,Iryna Zaporozhets, Cecilia Clementi, Christopher M. Bishop, Frank Noé Microsoft Research AI for Science (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-16634.txt.
Description
Yu Xie and a large team tackle rare-event sampling, the long-standing bottleneck in molecular dynamics, above all for biomolecules. Diffusion models such as BioEmu now draw free samples from tangled molecular distributions, cutting the cost of rare transitions. Yet a sampling snag stays when scores hinge on states that are rare at equilibrium, such as folding free energies.
The authors bring in enhanced diffusion sampling, which steers runs into rare zones while keeping thermodynamic scores unbiased. Careful steering builds biased sets, and exact reweighting restores equilibrium stats. Three codes embody it: UmbrellaDiff, MetaDiff, and ΔG-Diff. Tests on toy systems, protein folding maps, and folding free energies give fast, true scores within minutes to hours per system on graphics hardware.
Catalog record
| CREATOR | Yu Xie, Ludwig Winkler, Lixin Sun, Sarah Lewis, Adam E. FosterJosé Jiménez Luna, Tim Hempel, Michael Gastegger, Yaoyi Chen,Iryna Zaporozhets, Cecilia Clementi, Christopher M. Bishop, Frank Noé Microsoft Research AI for Science |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 55 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-16634