Hyperbolic Binary Neural Network

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Chen, Jun, Xiang, Jingyang, Huang, Tianxin, Zhao, Xiangrui, Liu, Yong
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866916553550725120
author Chen, Jun
Xiang, Jingyang
Huang, Tianxin
Zhao, Xiangrui
Liu, Yong
author_facet Chen, Jun
Xiang, Jingyang
Huang, Tianxin
Zhao, Xiangrui
Liu, Yong
contents Binary Neural Network (BNN) converts full-precision weights and activations into their extreme 1-bit counterparts, making it particularly suitable for deployment on lightweight mobile devices. While binary neural networks are typically formulated as a constrained optimization problem and optimized in the binarized space, general neural networks are formulated as an unconstrained optimization problem and optimized in the continuous space. This paper introduces the Hyperbolic Binary Neural Network (HBNN) by leveraging the framework of hyperbolic geometry to optimize the constrained problem. Specifically, we transform the constrained problem in hyperbolic space into an unconstrained one in Euclidean space using the Riemannian exponential map. On the other hand, we also propose the Exponential Parametrization Cluster (EPC) method, which, compared to the Riemannian exponential map, shrinks the segment domain based on a diffeomorphism. This approach increases the probability of weight flips, thereby maximizing the information gain in BNNs. Experimental results on CIFAR10, CIFAR100, and ImageNet classification datasets with VGGsmall, ResNet18, and ResNet34 models illustrate the superior performance of our HBNN over state-of-the-art methods.
format Preprint
id arxiv_https___arxiv_org_abs_2501_03471
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Hyperbolic Binary Neural Network
Chen, Jun
Xiang, Jingyang
Huang, Tianxin
Zhao, Xiangrui
Liu, Yong
Machine Learning
Computer Vision and Pattern Recognition
Binary Neural Network (BNN) converts full-precision weights and activations into their extreme 1-bit counterparts, making it particularly suitable for deployment on lightweight mobile devices. While binary neural networks are typically formulated as a constrained optimization problem and optimized in the binarized space, general neural networks are formulated as an unconstrained optimization problem and optimized in the continuous space. This paper introduces the Hyperbolic Binary Neural Network (HBNN) by leveraging the framework of hyperbolic geometry to optimize the constrained problem. Specifically, we transform the constrained problem in hyperbolic space into an unconstrained one in Euclidean space using the Riemannian exponential map. On the other hand, we also propose the Exponential Parametrization Cluster (EPC) method, which, compared to the Riemannian exponential map, shrinks the segment domain based on a diffeomorphism. This approach increases the probability of weight flips, thereby maximizing the information gain in BNNs. Experimental results on CIFAR10, CIFAR100, and ImageNet classification datasets with VGGsmall, ResNet18, and ResNet34 models illustrate the superior performance of our HBNN over state-of-the-art methods.
title Hyperbolic Binary Neural Network
topic Machine Learning
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2501.03471