Revisiting the problem of learning long-term dependencies in recurrent neural networks.

IF 6 1区 计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE
Liam Johnston, Vivak Patel, Yumian Cui, Prasanna Balaprakash
{"title":"Revisiting the problem of learning long-term dependencies in recurrent neural networks.","authors":"Liam Johnston, Vivak Patel, Yumian Cui, Prasanna Balaprakash","doi":"10.1016/j.neunet.2024.106887","DOIUrl":null,"url":null,"abstract":"<p><p>Recurrent neural networks (RNNs) are an important class of models for learning sequential behavior. However, training RNNs to learn long-term dependencies is a tremendously difficult task, and this difficulty is widely attributed to the vanishing and exploding gradient (VEG) problem. Since it was first characterized 30 years ago, the belief that if VEG occurs during optimization then RNNs learn long-term dependencies poorly has become a central tenet in the RNN literature and has been steadily cited as motivation for a wide variety of research advancements. In this work, we revisit and interrogate this belief using a large factorial experiment where more than 40,000 RNNs were trained, and provide evidence contradicting this belief. Motivated by these findings, we re-examine the original discussion that analyzed latching behavior in RNNs by way of hyperbolic attractors, and ultimately demonstrate that these dynamics do not fully capture the learned characteristics of RNNs. Our findings suggest that these models are fully capable of learning dynamics that do not correspond to hyperbolic attractors, and that the choice of hyper-parameters, namely learning rate, has a substantial impact on the likelihood of whether an RNN will be able to learn long-term dependencies.</p>","PeriodicalId":49763,"journal":{"name":"Neural Networks","volume":"183 ","pages":"106887"},"PeriodicalIF":6.0000,"publicationDate":"2024-11-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Neural Networks","FirstCategoryId":"94","ListUrlMain":"https://doi.org/10.1016/j.neunet.2024.106887","RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE","Score":null,"Total":0}
引用次数: 0

Abstract

Recurrent neural networks (RNNs) are an important class of models for learning sequential behavior. However, training RNNs to learn long-term dependencies is a tremendously difficult task, and this difficulty is widely attributed to the vanishing and exploding gradient (VEG) problem. Since it was first characterized 30 years ago, the belief that if VEG occurs during optimization then RNNs learn long-term dependencies poorly has become a central tenet in the RNN literature and has been steadily cited as motivation for a wide variety of research advancements. In this work, we revisit and interrogate this belief using a large factorial experiment where more than 40,000 RNNs were trained, and provide evidence contradicting this belief. Motivated by these findings, we re-examine the original discussion that analyzed latching behavior in RNNs by way of hyperbolic attractors, and ultimately demonstrate that these dynamics do not fully capture the learned characteristics of RNNs. Our findings suggest that these models are fully capable of learning dynamics that do not correspond to hyperbolic attractors, and that the choice of hyper-parameters, namely learning rate, has a substantial impact on the likelihood of whether an RNN will be able to learn long-term dependencies.

求助全文
约1分钟内获得全文 求助全文
来源期刊
Neural Networks
Neural Networks 工程技术-计算机:人工智能
CiteScore
13.90
自引率
7.70%
发文量
425
审稿时长
67 days
期刊介绍: Neural Networks is a platform that aims to foster an international community of scholars and practitioners interested in neural networks, deep learning, and other approaches to artificial intelligence and machine learning. Our journal invites submissions covering various aspects of neural networks research, from computational neuroscience and cognitive modeling to mathematical analyses and engineering applications. By providing a forum for interdisciplinary discussions between biology and technology, we aim to encourage the development of biologically-inspired artificial intelligence.
×
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