Adaptive Compression of the Latent Space in Variational Autoencoders

Result code in IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21730%2F24%3A00377215" target="_blank" >RIV/68407700:21730/24:00377215 - isvavai.cz</a>

Result on the web

<a href="https://doi.org/10.1007/978-3-031-72332-2_7" target="_blank" >https://doi.org/10.1007/978-3-031-72332-2_7</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1007/978-3-031-72332-2_7" target="_blank" >10.1007/978-3-031-72332-2_7</a>

Result language

angličtina

Original language name

Adaptive Compression of the Latent Space in Variational Autoencoders

Original language description

Variational Autoencoders (VAEs) are powerful generative models that have been widely used in various fields, including image and text generation. However, one of the known challenges in using VAEs is the model’s sensitivity to its hyperparameters, such as the latent space size. This paper presents a simple extension of VAEs for automatically determining the optimal latent space size during the training process by gradually decreasing the latent size through neuron removal and observ ing the model performance. The proposed method is compared to the traditional and computationally costly hyperparameter grid search and is shown to be significantly faster while still achieving the best optimal dimensionality on four image datasets. Furthermore, we show that the final performance as well as the speed of our method is comparable to training on the optimal latent size from scratch, and might thus serve as a convenient substitute for example in low-resource scenarios.

Czech name

—

Czech description

—

Type

D - Article in proceedings

CEP classification

—

OECD FORD branch

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Project

Result was created during the realization of more than one project. More information in the Projects tab.

Continuities

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Publication year

2024

Confidentiality

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Article name in the collection

Artificial Neural Networks and Machine Learning – ICANN 2024 33rd International Conference on Artificial Neural Networks, Lugano, Switzerland, September 17–20, 2024, Proceedings, Part IX

ISBN

978-3-031-72355-1

ISSN

0302-9743

e-ISSN

1611-3349

Number of pages

13

Pages from-to

89-101

Publisher name

Springer, Cham

Place of publication

—

Event location

Lugano-Viganello

Event date

Sep 17, 2024

Type of event by nationality

WRD - Celosvětová akce

UT code for WoS article

001331868600007