Saved in:
Bibliographic Details
Main Authors: McCallum, Sam, Foster, James
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2410.11648
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916588533317632
author McCallum, Sam
Foster, James
author_facet McCallum, Sam
Foster, James
contents Training Neural ODEs requires backpropagating through an ODE solve. The state-of-the-art backpropagation method is recursive checkpointing that balances recomputation with memory cost. Here, we introduce a class of algebraically reversible ODE solvers that significantly improve upon both the time and memory cost of recursive checkpointing. The reversible solvers presented calculate exact gradients, are high-order and numerically stable -- strictly improving on previous reversible architectures.
format Preprint
id arxiv_https___arxiv_org_abs_2410_11648
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Efficient, Accurate and Stable Gradients for Neural ODEs
McCallum, Sam
Foster, James
Machine Learning
Training Neural ODEs requires backpropagating through an ODE solve. The state-of-the-art backpropagation method is recursive checkpointing that balances recomputation with memory cost. Here, we introduce a class of algebraically reversible ODE solvers that significantly improve upon both the time and memory cost of recursive checkpointing. The reversible solvers presented calculate exact gradients, are high-order and numerically stable -- strictly improving on previous reversible architectures.
title Efficient, Accurate and Stable Gradients for Neural ODEs
topic Machine Learning
url https://arxiv.org/abs/2410.11648