Denoising Multi-Beta VAE: Representation Learning for Disentanglement and Generation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Uppal, Anshuk, Takida, Yuhta, Lai, Chieh-Hsin, Mitsufuji, Yuki
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912472953257984
author Uppal, Anshuk
Takida, Yuhta
Lai, Chieh-Hsin
Mitsufuji, Yuki
author_facet Uppal, Anshuk
Takida, Yuhta
Lai, Chieh-Hsin
Mitsufuji, Yuki
contents Disentangled and interpretable latent representations in generative models typically come at the cost of generation quality. The $β$-VAE framework introduces a hyperparameter $β$ to balance disentanglement and reconstruction quality, where setting $β> 1$ introduces an information bottleneck that favors disentanglement over sharp, accurate reconstructions. To address this trade-off, we propose a novel generative modeling framework that leverages a range of $β$ values to learn multiple corresponding latent representations. First, we obtain a slew of representations by training a single variational autoencoder (VAE), with a new loss function that controls the information retained in each latent representation such that the higher $β$ value prioritize disentanglement over reconstruction fidelity. We then, introduce a non-linear diffusion model that smoothly transitions latent representations corresponding to different $β$ values. This model denoises towards less disentangled and more informative representations, ultimately leading to (almost) lossless representations, enabling sharp reconstructions. Furthermore, our model supports sample generation without input images, functioning as a standalone generative model. We evaluate our framework in terms of both disentanglement and generation quality. Additionally, we observe smooth transitions in the latent spaces with respect to changes in $β$, facilitating consistent manipulation of generated outputs.
format Preprint
id arxiv_https___arxiv_org_abs_2507_06613
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Denoising Multi-Beta VAE: Representation Learning for Disentanglement and Generation
Uppal, Anshuk
Takida, Yuhta
Lai, Chieh-Hsin
Mitsufuji, Yuki
Machine Learning
Artificial Intelligence
Computer Vision and Pattern Recognition
Disentangled and interpretable latent representations in generative models typically come at the cost of generation quality. The $β$-VAE framework introduces a hyperparameter $β$ to balance disentanglement and reconstruction quality, where setting $β> 1$ introduces an information bottleneck that favors disentanglement over sharp, accurate reconstructions. To address this trade-off, we propose a novel generative modeling framework that leverages a range of $β$ values to learn multiple corresponding latent representations. First, we obtain a slew of representations by training a single variational autoencoder (VAE), with a new loss function that controls the information retained in each latent representation such that the higher $β$ value prioritize disentanglement over reconstruction fidelity. We then, introduce a non-linear diffusion model that smoothly transitions latent representations corresponding to different $β$ values. This model denoises towards less disentangled and more informative representations, ultimately leading to (almost) lossless representations, enabling sharp reconstructions. Furthermore, our model supports sample generation without input images, functioning as a standalone generative model. We evaluate our framework in terms of both disentanglement and generation quality. Additionally, we observe smooth transitions in the latent spaces with respect to changes in $β$, facilitating consistent manipulation of generated outputs.
title Denoising Multi-Beta VAE: Representation Learning for Disentanglement and Generation
topic Machine Learning
Artificial Intelligence
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2507.06613