by V. R. Somnath, M. Pariset, Y. P. Hsieh, M. R. Martinez, A. Krause, C. Bunne

Abstract:

Diffusion Schrödinger bridges (DSB) have recently emerged as a powerful framework for recovering stochastic dynamics via their marginal observations at different time points. Despite numerous successful applications, existing algorithms for solving DSBs have so far failed to utilize the structure of aligned data, which naturally arises in many biological phenomena. In this paper, we propose a novel algorithmic framework that, for the first time, solves DSBs while respecting the data alignment. Our approach hinges on a combination of two decades-old ideas: The classical Schrödinger bridge theory and Doob's $h$-transform. Compared to prior methods, our approach leads to a simpler training procedure with lower variance, which we further augment with principled regularization schemes. This ultimately leads to sizeable improvements across experiments on synthetic and real data, including the tasks of rigid protein docking and temporal evolution of cellular differentiation processes.

Reference:

Aligned Diffusion Schrödinger Bridges V. R. Somnath, M. Pariset, Y. P. Hsieh, M. R. Martinez, A. Krause, C. BunneIn Conference on Uncertainty in Artificial Intelligence (UAI), 2023

Bibtex Entry:

@inproceedings{somnath2023aligneddinger bridges (DSB) have recently emerged as a powerful framework for recovering stochastic dynamics via their marginal observations at different time points. Despite numerous successful applications, existing algorithms for solving DSBs have so far failed to utilize the structure of aligned data, which naturally arises in many biological phenomena. In this paper, we propose a novel algorithmic framework that, for the first time, solves DSBs while respecting the data alignment. Our approach hinges on a combination of two decades-old ideas: The classical Schr{\"o}dinger bridge theory and Doob's $h$-transform. Compared to prior methods, our approach leads to a simpler training procedure with lower variance, which we further augment with principled regularization schemes. This ultimately leads to sizeable improvements across experiments on synthetic and real data, including the tasks of rigid protein docking and temporal evolution of cellular differentiation processes.},
	author = {Vignesh Ram Somnath and Matteo Pariset and Ya-Ping Hsieh and Maria Rodriguez Martinez and Andreas Krause and Charlotte Bunne},
	booktitle = {Conference on Uncertainty in Artificial Intelligence (UAI)},
	month = {August},
	title = {{Aligned Diffusion Schr{\"o}dinger Bridges}},
	year = {2023}}