Geometries of information and sampling free energies
Gavin E Crooks
2025 Invited Talk
in
Workshop: Frontiers in Probabilistic Inference: Learning meets Sampling
in
Workshop: Frontiers in Probabilistic Inference: Learning meets Sampling
Abstract
The efficiency of estimating free energies can be understood in terms of the geometry of the informational and dynamical space underlying the sampling problem. There are different levels of description, reflecting different constraints and approximations: I'll discuss the relation between Fisher, thermodynamic, linear response, and Weinhold geometries.
Speaker
Video
Chat is not available.
Successful Page Load