Rank Reduction Autoencoders

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Mounayer, Jad, Rodriguez, Sebastian, Ghnatios, Chady, Farhat, Charbel, Chinesta, Francisco
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866917967435923456
author Mounayer, Jad
Rodriguez, Sebastian
Ghnatios, Chady
Farhat, Charbel
Chinesta, Francisco
author_facet Mounayer, Jad
Rodriguez, Sebastian
Ghnatios, Chady
Farhat, Charbel
Chinesta, Francisco
contents The choice of an appropriate bottleneck dimension and the application of effective regularization are both essential for Autoencoders to learn meaningful representations from unlabeled data. In this paper, we introduce a new class of deterministic autoencoders, Rank Reduction Autoencoders (RRAEs), which regularize their latent spaces by employing a truncated singular value decomposition (SVD) during training. In RRAEs, the bottleneck is defined by the rank of the latent matrix, thereby alleviating the dependence of the encoder/decoder architecture on the bottleneck size. This approach enabled us to propose an adaptive algorithm (aRRAEs) that efficiently determines the optimal bottleneck size during training. We empirically demonstrate that both RRAEs and aRRAEs are stable, scalable, and reliable, as they do not introduce any additional training hyperparameters. We evaluate our proposed architecture on a synthetic data set, as well as on MNIST, Fashion MNIST, and CelebA. Our results show that RRAEs offer several advantages over Vanilla AEs with both large and small latent spaces, and outperform other regularizing AE architectures.
format Preprint
id arxiv_https___arxiv_org_abs_2405_13980
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Rank Reduction Autoencoders
Mounayer, Jad
Rodriguez, Sebastian
Ghnatios, Chady
Farhat, Charbel
Chinesta, Francisco
Machine Learning
The choice of an appropriate bottleneck dimension and the application of effective regularization are both essential for Autoencoders to learn meaningful representations from unlabeled data. In this paper, we introduce a new class of deterministic autoencoders, Rank Reduction Autoencoders (RRAEs), which regularize their latent spaces by employing a truncated singular value decomposition (SVD) during training. In RRAEs, the bottleneck is defined by the rank of the latent matrix, thereby alleviating the dependence of the encoder/decoder architecture on the bottleneck size. This approach enabled us to propose an adaptive algorithm (aRRAEs) that efficiently determines the optimal bottleneck size during training. We empirically demonstrate that both RRAEs and aRRAEs are stable, scalable, and reliable, as they do not introduce any additional training hyperparameters. We evaluate our proposed architecture on a synthetic data set, as well as on MNIST, Fashion MNIST, and CelebA. Our results show that RRAEs offer several advantages over Vanilla AEs with both large and small latent spaces, and outperform other regularizing AE architectures.
title Rank Reduction Autoencoders
topic Machine Learning
url https://arxiv.org/abs/2405.13980