We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

stat.ML

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Statistics > Machine Learning

Title: Smooth Normalizing Flows

Abstract: Normalizing flows are a promising tool for modeling probability distributions in physical systems. While state-of-the-art flows accurately approximate distributions and energies, applications in physics additionally require smooth energies to compute forces and higher-order derivatives. Furthermore, such densities are often defined on non-trivial topologies. A recent example are Boltzmann Generators for generating 3D-structures of peptides and small proteins. These generative models leverage the space of internal coordinates (dihedrals, angles, and bonds), which is a product of hypertori and compact intervals. In this work, we introduce a class of smooth mixture transformations working on both compact intervals and hypertori. Mixture transformations employ root-finding methods to invert them in practice, which has so far prevented bi-directional flow training. To this end, we show that parameter gradients and forces of such inverses can be computed from forward evaluations via the inverse function theorem. We demonstrate two advantages of such smooth flows: they allow training by force matching to simulation data and can be used as potentials in molecular dynamics simulations.
Comments: Neural Information Proceessing Systems (NeurIPS) 2021
Subjects: Machine Learning (stat.ML); Machine Learning (cs.LG); Chemical Physics (physics.chem-ph)
Cite as: arXiv:2110.00351 [stat.ML]
  (or arXiv:2110.00351v2 [stat.ML] for this version)

Submission history

From: Jonas Köhler [view email]
[v1] Fri, 1 Oct 2021 12:27:14 GMT (5454kb,D)
[v2] Tue, 30 Nov 2021 18:24:05 GMT (7629kb,D)

Link back to: arXiv, form interface, contact.