Saved in:
Bibliographic Details
Main Authors: Baker, Elizabeth L., Schauer, Moritz, Sommer, Stefan
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2407.15455
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915195638513664
author Baker, Elizabeth L.
Schauer, Moritz
Sommer, Stefan
author_facet Baker, Elizabeth L.
Schauer, Moritz
Sommer, Stefan
contents We propose a new algorithm for learning bridged diffusion processes using score-matching methods. Our method relies on reversing the dynamics of the forward process and using this to learn a score function, which, via Doob's $h$-transform, yields a bridged diffusion process; that is, a process conditioned on an endpoint. In contrast to prior methods, we learn the score term $\nabla_x \log p(t, x; T, y)$ directly, for given $t, y$, completely avoiding first learning a time-reversal. We compare the performance of our algorithm with existing methods and see that it outperforms using the (learned) time-reversals to learn the score term. The code can be found at https://github.com/libbylbaker/forward_bridge.
format Preprint
id arxiv_https___arxiv_org_abs_2407_15455
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Score matching for bridges without learning time-reversals
Baker, Elizabeth L.
Schauer, Moritz
Sommer, Stefan
Machine Learning
Probability
We propose a new algorithm for learning bridged diffusion processes using score-matching methods. Our method relies on reversing the dynamics of the forward process and using this to learn a score function, which, via Doob's $h$-transform, yields a bridged diffusion process; that is, a process conditioned on an endpoint. In contrast to prior methods, we learn the score term $\nabla_x \log p(t, x; T, y)$ directly, for given $t, y$, completely avoiding first learning a time-reversal. We compare the performance of our algorithm with existing methods and see that it outperforms using the (learned) time-reversals to learn the score term. The code can be found at https://github.com/libbylbaker/forward_bridge.
title Score matching for bridges without learning time-reversals
topic Machine Learning
Probability
url https://arxiv.org/abs/2407.15455