Efficient Model Compression for Bayesian Neural Networks

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Saha, Diptarka, Liu, Zihe, Liang, Feng
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866910680426217472
author Saha, Diptarka
Liu, Zihe
Liang, Feng
author_facet Saha, Diptarka
Liu, Zihe
Liang, Feng
contents Model Compression has drawn much attention within the deep learning community recently. Compressing a dense neural network offers many advantages including lower computation cost, deployability to devices of limited storage and memories, and resistance to adversarial attacks. This may be achieved via weight pruning or fully discarding certain input features. Here we demonstrate a novel strategy to emulate principles of Bayesian model selection in a deep learning setup. Given a fully connected Bayesian neural network with spike-and-slab priors trained via a variational algorithm, we obtain the posterior inclusion probability for every node that typically gets lost. We employ these probabilities for pruning and feature selection on a host of simulated and real-world benchmark data and find evidence of better generalizability of the pruned model in all our experiments.
format Preprint
id arxiv_https___arxiv_org_abs_2411_00273
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Efficient Model Compression for Bayesian Neural Networks
Saha, Diptarka
Liu, Zihe
Liang, Feng
Machine Learning
Applications
Model Compression has drawn much attention within the deep learning community recently. Compressing a dense neural network offers many advantages including lower computation cost, deployability to devices of limited storage and memories, and resistance to adversarial attacks. This may be achieved via weight pruning or fully discarding certain input features. Here we demonstrate a novel strategy to emulate principles of Bayesian model selection in a deep learning setup. Given a fully connected Bayesian neural network with spike-and-slab priors trained via a variational algorithm, we obtain the posterior inclusion probability for every node that typically gets lost. We employ these probabilities for pruning and feature selection on a host of simulated and real-world benchmark data and find evidence of better generalizability of the pruned model in all our experiments.
title Efficient Model Compression for Bayesian Neural Networks
topic Machine Learning
Applications
url https://arxiv.org/abs/2411.00273