Saved in:
Bibliographic Details
Main Authors: Kim, Bum Jun, Kim, Sang Woo
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2409.16630
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913517311885312
author Kim, Bum Jun
Kim, Sang Woo
author_facet Kim, Bum Jun
Kim, Sang Woo
contents Regularization of deep neural networks has been an important issue to achieve higher generalization performance without overfitting problems. Although the popular method of Dropout provides a regularization effect, it causes inconsistent properties in the output, which may degrade the performance of deep neural networks. In this study, we propose a new module called stochastic average pooling, which incorporates Dropout-like stochasticity in pooling. We describe the properties of stochastic subsampling and average pooling and leverage them to design a module without any inconsistency problem. The stochastic average pooling achieves a regularization effect without any potential performance degradation due to the inconsistency issue and can easily be plugged into existing architectures of deep neural networks. Experiments demonstrate that replacing existing average pooling with stochastic average pooling yields consistent improvements across a variety of tasks, datasets, and models.
format Preprint
id arxiv_https___arxiv_org_abs_2409_16630
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Stochastic Subsampling With Average Pooling
Kim, Bum Jun
Kim, Sang Woo
Machine Learning
Artificial Intelligence
Computer Vision and Pattern Recognition
Regularization of deep neural networks has been an important issue to achieve higher generalization performance without overfitting problems. Although the popular method of Dropout provides a regularization effect, it causes inconsistent properties in the output, which may degrade the performance of deep neural networks. In this study, we propose a new module called stochastic average pooling, which incorporates Dropout-like stochasticity in pooling. We describe the properties of stochastic subsampling and average pooling and leverage them to design a module without any inconsistency problem. The stochastic average pooling achieves a regularization effect without any potential performance degradation due to the inconsistency issue and can easily be plugged into existing architectures of deep neural networks. Experiments demonstrate that replacing existing average pooling with stochastic average pooling yields consistent improvements across a variety of tasks, datasets, and models.
title Stochastic Subsampling With Average Pooling
topic Machine Learning
Artificial Intelligence
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2409.16630