用于分词的迭代扩展卷积神经网络

IF 0.7 4区 计算机科学 Q4 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE
H. He, X. Yang, L. Wu, G. Wang
{"title":"用于分词的迭代扩展卷积神经网络","authors":"H. He, X. Yang, L. Wu, G. Wang","doi":"10.14311/NNW.2020.30.022","DOIUrl":null,"url":null,"abstract":"The latest development of neural word segmentation is governed by bi-directional Long Short-Term Memory Networks (Bi-LSTMs) that utilize Recurrent Neural Networks (RNNs) as standard sequence tagging models, resulting in expressive and accurate performance on large-scale dataset. However, RNNs are not adapted to fully exploit the parallelism capability of Graphics Processing Unit (GPU), limiting their computational efficiency in both learning and inferring phases. This paper proposes a novel approach adopting Iterated Dilated Convolutional Neural Networks (ID-CNNs) to supersede Bi-LSTMs for faster computation while retaining accuracy. Our implementation has achieved state-of-the-art result on SIGHAN Bakeoff 2005 datasets. Extensive experiments showed that our approach with ID-CNNs enables 3× training time speedups with no accuracy loss, achieving better accuracy compared to the prevailing Bi-LSTMs. Source code and corpora of this paper have been made publicly available on GitHub.","PeriodicalId":49765,"journal":{"name":"Neural Network World","volume":"30 1","pages":"333-346"},"PeriodicalIF":0.7000,"publicationDate":"2020-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"3","resultStr":"{\"title\":\"Iterated dilated convolutional neural networks for word segmentation\",\"authors\":\"H. He, X. Yang, L. Wu, G. Wang\",\"doi\":\"10.14311/NNW.2020.30.022\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"The latest development of neural word segmentation is governed by bi-directional Long Short-Term Memory Networks (Bi-LSTMs) that utilize Recurrent Neural Networks (RNNs) as standard sequence tagging models, resulting in expressive and accurate performance on large-scale dataset. However, RNNs are not adapted to fully exploit the parallelism capability of Graphics Processing Unit (GPU), limiting their computational efficiency in both learning and inferring phases. This paper proposes a novel approach adopting Iterated Dilated Convolutional Neural Networks (ID-CNNs) to supersede Bi-LSTMs for faster computation while retaining accuracy. Our implementation has achieved state-of-the-art result on SIGHAN Bakeoff 2005 datasets. Extensive experiments showed that our approach with ID-CNNs enables 3× training time speedups with no accuracy loss, achieving better accuracy compared to the prevailing Bi-LSTMs. Source code and corpora of this paper have been made publicly available on GitHub.\",\"PeriodicalId\":49765,\"journal\":{\"name\":\"Neural Network World\",\"volume\":\"30 1\",\"pages\":\"333-346\"},\"PeriodicalIF\":0.7000,\"publicationDate\":\"2020-01-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"3\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Neural Network World\",\"FirstCategoryId\":\"94\",\"ListUrlMain\":\"https://doi.org/10.14311/NNW.2020.30.022\",\"RegionNum\":4,\"RegionCategory\":\"计算机科学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q4\",\"JCRName\":\"COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Neural Network World","FirstCategoryId":"94","ListUrlMain":"https://doi.org/10.14311/NNW.2020.30.022","RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q4","JCRName":"COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE","Score":null,"Total":0}
引用次数: 3

摘要

神经分词的最新发展是双向长短期记忆网络(Bi-LSTMs),它利用递归神经网络(RNNs)作为标准的序列标记模型,从而在大规模数据集上具有表达性和准确性。然而,rnn不能充分利用图形处理单元(GPU)的并行能力,限制了其在学习和推理阶段的计算效率。本文提出了一种采用迭代扩展卷积神经网络(id - cnn)取代bi - lstm的新方法,以提高计算速度,同时保持精度。我们的实现在SIGHAN Bakeoff 2005数据集上取得了最先进的结果。大量的实验表明,我们的id - cnn方法可以在没有精度损失的情况下实现3倍的训练时间加速,与目前流行的Bi-LSTMs相比,可以获得更好的精度。本文的源代码和语料库已在GitHub上公开发布。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
Iterated dilated convolutional neural networks for word segmentation
The latest development of neural word segmentation is governed by bi-directional Long Short-Term Memory Networks (Bi-LSTMs) that utilize Recurrent Neural Networks (RNNs) as standard sequence tagging models, resulting in expressive and accurate performance on large-scale dataset. However, RNNs are not adapted to fully exploit the parallelism capability of Graphics Processing Unit (GPU), limiting their computational efficiency in both learning and inferring phases. This paper proposes a novel approach adopting Iterated Dilated Convolutional Neural Networks (ID-CNNs) to supersede Bi-LSTMs for faster computation while retaining accuracy. Our implementation has achieved state-of-the-art result on SIGHAN Bakeoff 2005 datasets. Extensive experiments showed that our approach with ID-CNNs enables 3× training time speedups with no accuracy loss, achieving better accuracy compared to the prevailing Bi-LSTMs. Source code and corpora of this paper have been made publicly available on GitHub.
求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
Neural Network World
Neural Network World 工程技术-计算机:人工智能
CiteScore
1.80
自引率
0.00%
发文量
0
审稿时长
12 months
期刊介绍: Neural Network World is a bimonthly journal providing the latest developments in the field of informatics with attention mainly devoted to the problems of: brain science, theory and applications of neural networks (both artificial and natural), fuzzy-neural systems, methods and applications of evolutionary algorithms, methods of parallel and mass-parallel computing, problems of soft-computing, methods of 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