Guiding Evolutionary AutoEncoder Training with Activation-Based Pruning Operators

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Jorgensen, Steven, Hemberg, Erik, Toutouh, Jamal, O'Reilly, Una-May
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866918013355163648
author Jorgensen, Steven
Hemberg, Erik
Toutouh, Jamal
O'Reilly, Una-May
author_facet Jorgensen, Steven
Hemberg, Erik
Toutouh, Jamal
O'Reilly, Una-May
contents This study explores a novel approach to neural network pruning using evolutionary computation, focusing on simultaneously pruning the encoder and decoder of an autoencoder. We introduce two new mutation operators that use layer activations to guide weight pruning. Our findings reveal that one of these activation-informed operators outperforms random pruning, resulting in more efficient autoencoders with comparable performance to canonically trained models. Prior work has established that autoencoder training is effective and scalable with a spatial coevolutionary algorithm that cooperatively coevolves a population of encoders with a population of decoders, rather than one autoencoder. We evaluate how the same activity-guided mutation operators transfer to this context. We find that random pruning is better than guided pruning, in the coevolutionary setting. This suggests activation-based guidance proves more effective in low-dimensional pruning environments, where constrained sample spaces can lead to deviations from true uniformity in randomization. Conversely, population-driven strategies enhance robustness by expanding the total pruning dimensionality, achieving statistically uniform randomness that better preserves system dynamics. We experiment with pruning according to different schedules and present best combinations of operator and schedule for the canonical and coevolving populations cases.
format Preprint
id arxiv_https___arxiv_org_abs_2505_05138
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Guiding Evolutionary AutoEncoder Training with Activation-Based Pruning Operators
Jorgensen, Steven
Hemberg, Erik
Toutouh, Jamal
O'Reilly, Una-May
Neural and Evolutionary Computing
Artificial Intelligence
This study explores a novel approach to neural network pruning using evolutionary computation, focusing on simultaneously pruning the encoder and decoder of an autoencoder. We introduce two new mutation operators that use layer activations to guide weight pruning. Our findings reveal that one of these activation-informed operators outperforms random pruning, resulting in more efficient autoencoders with comparable performance to canonically trained models. Prior work has established that autoencoder training is effective and scalable with a spatial coevolutionary algorithm that cooperatively coevolves a population of encoders with a population of decoders, rather than one autoencoder. We evaluate how the same activity-guided mutation operators transfer to this context. We find that random pruning is better than guided pruning, in the coevolutionary setting. This suggests activation-based guidance proves more effective in low-dimensional pruning environments, where constrained sample spaces can lead to deviations from true uniformity in randomization. Conversely, population-driven strategies enhance robustness by expanding the total pruning dimensionality, achieving statistically uniform randomness that better preserves system dynamics. We experiment with pruning according to different schedules and present best combinations of operator and schedule for the canonical and coevolving populations cases.
title Guiding Evolutionary AutoEncoder Training with Activation-Based Pruning Operators
topic Neural and Evolutionary Computing
Artificial Intelligence
url https://arxiv.org/abs/2505.05138