Salvato in:
Dettagli Bibliografici
Autore principale: Golbert, Simon
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:https://arxiv.org/abs/2505.03791
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866910929910759424
author Golbert, Simon
author_facet Golbert, Simon
contents Boolean neural networks offer hardware-efficient alternatives to real-valued models. While quantization is common, purely Boolean training remains underexplored. We present a practical method for purely Boolean backpropagation for networks based on a single specific gate we chose, operating directly in Boolean algebra involving no numerics. Initial experiments confirm its feasibility.
format Preprint
id arxiv_https___arxiv_org_abs_2505_03791
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Practical Boolean Backpropagation
Golbert, Simon
Machine Learning
Artificial Intelligence
Boolean neural networks offer hardware-efficient alternatives to real-valued models. While quantization is common, purely Boolean training remains underexplored. We present a practical method for purely Boolean backpropagation for networks based on a single specific gate we chose, operating directly in Boolean algebra involving no numerics. Initial experiments confirm its feasibility.
title Practical Boolean Backpropagation
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2505.03791