Physics-informed Variational Autoencoders for Improved Robustness to Environmental Factors of Variation

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Thoreau, Romain, Risser, Laurent, Achard, Véronique, Berthelot, Béatrice, Briottet, Xavier
Natura: Preprint
Pubblicazione: 2022
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866915172777459712
author Thoreau, Romain
Risser, Laurent
Achard, Véronique
Berthelot, Béatrice
Briottet, Xavier
author_facet Thoreau, Romain
Risser, Laurent
Achard, Véronique
Berthelot, Béatrice
Briottet, Xavier
contents The combination of machine learning models with physical models is a recent research path to learn robust data representations. In this paper, we introduce p$^3$VAE, a variational autoencoder that integrates prior physical knowledge about the latent factors of variation that are related to the data acquisition conditions. p$^3$VAE combines standard neural network layers with non-trainable physics layers in order to partially ground the latent space to physical variables. We introduce a semi-supervised learning algorithm that strikes a balance between the machine learning part and the physics part. Experiments on simulated and real data sets demonstrate the benefits of our framework against competing physics-informed and conventional machine learning models, in terms of extrapolation capabilities and interpretability. In particular, we show that p$^3$VAE naturally has interesting disentanglement capabilities. Our code and data have been made publicly available at https://github.com/Romain3Ch216/p3VAE.
format Preprint
id arxiv_https___arxiv_org_abs_2210_10418
institution arXiv
publishDate 2022
record_format arxiv
spellingShingle Physics-informed Variational Autoencoders for Improved Robustness to Environmental Factors of Variation
Thoreau, Romain
Risser, Laurent
Achard, Véronique
Berthelot, Béatrice
Briottet, Xavier
Computer Vision and Pattern Recognition
Machine Learning
68T45
I.2.6; I.2.10
The combination of machine learning models with physical models is a recent research path to learn robust data representations. In this paper, we introduce p$^3$VAE, a variational autoencoder that integrates prior physical knowledge about the latent factors of variation that are related to the data acquisition conditions. p$^3$VAE combines standard neural network layers with non-trainable physics layers in order to partially ground the latent space to physical variables. We introduce a semi-supervised learning algorithm that strikes a balance between the machine learning part and the physics part. Experiments on simulated and real data sets demonstrate the benefits of our framework against competing physics-informed and conventional machine learning models, in terms of extrapolation capabilities and interpretability. In particular, we show that p$^3$VAE naturally has interesting disentanglement capabilities. Our code and data have been made publicly available at https://github.com/Romain3Ch216/p3VAE.
title Physics-informed Variational Autoencoders for Improved Robustness to Environmental Factors of Variation
topic Computer Vision and Pattern Recognition
Machine Learning
68T45
I.2.6; I.2.10
url https://arxiv.org/abs/2210.10418