Provably Convergent Schrödinger Bridge with Applications to Probabilistic Time Series Imputation
Yu Chen, Wei Deng, Shikai Fang, Fengpei Li, Tianjiao Nicole Yang, Yikai Zhang, Kashif Rasul, Shandian Zhe
Abstract
The Schrödinger bridge problem (SBP) is gaining increasing attention in generative modeling and showing promising potential even in comparison with the score-based generative models (SGMs). SBP can be interpreted as an entropy-regularized optimal transport problem, which conducts projections onto every other marginal alternatingly. However, in practice, only approximated projections are accessible and their convergence is not well understood. To fill this gap, we present a first convergence analysis of the Schrödinger bridge algorithm based on approximated projections. As for its practical applications, we apply SBP to probabilistic time series imputation by generating missing values conditioned on observed data. We show that optimizing the transport cost improves the performance and the proposed algorithm achieves the state-of-the-art result in healthcare and environmental data while exhibiting the advantage of exploring both temporal and feature patterns in probabilistic time series imputation.
BibTeX
@inproceedings{icml2023_provablyconverge,
title = {Provably Convergent Schrödinger Bridge with Applications to Probabilistic Time Series Imputation},
author = {Yu Chen and Wei Deng and Shikai Fang and Fengpei Li and Tianjiao Nicole Yang and Yikai Zhang and Kashif Rasul and Shandian Zhe and Anderson Schneider and Yuriy Nevmyvaka},
booktitle = {ICML 2023},
year = {2023}
}