Convergence rates for an Adaptive Biasing Potential scheme from a Wasserstein optimization perspective
Tony Lelièvre, Xuyang Lin, Pierre Monmarché · 2025
Read Convergence rates for an Adaptive Biasing Potential scheme from a Wasserstein optimization perspective by Tony Lelièvre, Xuyang Lin, Pierre Monmarché (2025) 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-2501-17979.txt.
Introduction
Tony Lelièvre, Xuyang Lin, and Pierre Monmarché prove the first exponential convergence result for a family of simulation methods used to explore molecular energy landscapes, in a 2025 paper. Adaptive biasing methods, including metadynamics and the adaptive biasing force, add a shifting push that helps simulations cross energy barriers. They have worked well in practice for decades, but proving how fast they settle is hard, especially for the realistic kinetic Langevin dynamics, which is non-reversible.
Working in an idealized setting, the authors recast the algorithm as a flow over probability distributions, a mean-field picture. The key insight is that the process amounts to gradient descent of a well-chosen functional on that distribution space. From that viewpoint the exponential rate follows. The authors are explicit that the setting is idealized, but the result gives this widely used family its first firm convergence footing of this strength.
Catalog record
| CREATOR | Tony Lelièvre, Xuyang Lin, Pierre Monmarché |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 186 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2501-17979