STP‐CNN: Selection of transfer parameters in convolutional neural networks

IF 3 4区 计算机科学 Q2 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE
Expert Systems Pub Date : 2024-09-12 DOI:10.1111/exsy.13728
Otmane Mallouk, Nour‐Eddine Joudar, Mohamed Ettaouil
{"title":"STP‐CNN: Selection of transfer parameters in convolutional neural networks","authors":"Otmane Mallouk, Nour‐Eddine Joudar, Mohamed Ettaouil","doi":"10.1111/exsy.13728","DOIUrl":null,"url":null,"abstract":"Nowadays, transfer learning has shown promising results in many applications. However, most deep transfer learning methods such as <jats:italic>parameter sharing</jats:italic> and <jats:italic>fine‐tuning</jats:italic> are still suffering from the lack of parameters transmission strategy. In this paper, we propose a new optimization model for parameter‐based transfer learning in convolutional neural networks named STP‐CNN. Indeed, we propose a Lasso transfer model supported by a regularization term that controls transferability. Moreover, we opt for the proximal gradient descent method to solve the proposed model. The suggested technique allows, under certain conditions, to control exactly which parameters, in each convolutional layer of the source network, which will be used directly or adjusted in the target network. Several experiments prove the performance of our model in locating the transferable parameters as well as improving the data classification.","PeriodicalId":51053,"journal":{"name":"Expert Systems","volume":null,"pages":null},"PeriodicalIF":3.0000,"publicationDate":"2024-09-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Expert Systems","FirstCategoryId":"94","ListUrlMain":"https://doi.org/10.1111/exsy.13728","RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE","Score":null,"Total":0}
引用次数: 0

Abstract

Nowadays, transfer learning has shown promising results in many applications. However, most deep transfer learning methods such as parameter sharing and fine‐tuning are still suffering from the lack of parameters transmission strategy. In this paper, we propose a new optimization model for parameter‐based transfer learning in convolutional neural networks named STP‐CNN. Indeed, we propose a Lasso transfer model supported by a regularization term that controls transferability. Moreover, we opt for the proximal gradient descent method to solve the proposed model. The suggested technique allows, under certain conditions, to control exactly which parameters, in each convolutional layer of the source network, which will be used directly or adjusted in the target network. Several experiments prove the performance of our model in locating the transferable parameters as well as improving the data classification.
STP-CNN:选择卷积神经网络中的传递参数
如今,迁移学习在许多应用中都取得了可喜的成果。然而,大多数深度迁移学习方法,如参数共享和微调,仍存在缺乏参数传输策略的问题。在本文中,我们为卷积神经网络中基于参数的迁移学习提出了一种新的优化模型,命名为 STP-CNN。事实上,我们提出了一种 Lasso 转移模型,该模型由一个控制可转移性的正则化项支持。此外,我们还选择了近似梯度下降法来求解所提出的模型。在某些条件下,所建议的技术可以精确控制源网络每个卷积层中的参数,这些参数将直接用于目标网络或在目标网络中进行调整。一些实验证明了我们的模型在定位可转移参数和改进数据分类方面的性能。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
Expert Systems
Expert Systems 工程技术-计算机:理论方法
CiteScore
7.40
自引率
6.10%
发文量
266
审稿时长
24 months
期刊介绍: Expert Systems: The Journal of Knowledge Engineering publishes papers dealing with all aspects of knowledge engineering, including individual methods and techniques in knowledge acquisition and representation, and their application in the construction of systems – including expert systems – based thereon. Detailed scientific evaluation is an essential part of any paper. As well as traditional application areas, such as Software and Requirements Engineering, Human-Computer Interaction, and Artificial Intelligence, we are aiming at the new and growing markets for these technologies, such as Business, Economy, Market Research, and Medical and Health Care. The shift towards this new focus will be marked by a series of special issues covering hot and emergent topics.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信