Predictive variational autoencoder for learning robust representations of time-series data

Wang, Julia H, Tsin, Dexter, Engel, Tatiana A (December 2023) Predictive variational autoencoder for learning robust representations of time-series data. arXiv. (Submitted)

PDF nihpp-2312.06932v1.pdf - Submitted Version Available under License Creative Commons Attribution Non-commercial. Download (29MB)

Abstract

Variational autoencoders (VAEs) have been used extensively to discover low-dimensional latent factors governing neural activity and animal behavior. However, without careful model selection, the uncovered latent factors may reflect noise in the data rather than true underlying features, rendering such representations unsuitable for scientific interpretation. Existing solutions to this problem involve introducing additional measured variables or data augmentations specific to a particular data type. We propose a VAE architecture that predicts the next point in time and show that it mitigates the learning of spurious features. In addition, we introduce a model selection metric based on smoothness over time in the latent space. We show that together these two constraints on VAEs to be smooth over time produce robust latent representations and faithfully recover latent factors on synthetic datasets.

Item Type:	Paper
Subjects:	bioinformatics bioinformatics > computational biology > algorithms bioinformatics > computational biology
CSHL Authors:	Wang, Julia Engel, Tatiana
Communities:	CSHL labs > Engel lab School of Biological Sciences > Publications
SWORD Depositor:	CSHL Elements
Depositing User:	CSHL Elements
Date:	12 December 2023
Date Deposited:	23 Jan 2024 16:49
Last Modified:	29 Feb 2024 18:56
PMCID:	PMC10760197
URI:	https://repository.cshl.edu/id/eprint/41417

Actions (login required)

Administrator's edit/view item