Neural Architectures for Named Entity Recognition. Lample, G., Ballesteros, M., Subramanian, S., Kawakami, K., & Dyer, C. arXiv:1603.01360 [cs], March, 2016. 🏷️ /unread、Computer Science - Computation and Language![Neural Architectures for Named Entity Recognition [link]](https://bibbase.org/img/filetypes/link.svg "link")
Paper abstract bibtex State-of-the-art named entity recognition systems rely heavily on hand-crafted features and domain-specific knowledge in order to learn effectively from the small, supervised training corpora that are available. In this paper, we introduce two new neural architectures—one based on bidirectional LSTMs and conditional random fields, and the other that constructs and labels segments using a transition-based approach inspired by shift-reduce parsers. Our models rely on two sources of information about words: character-based word representations learned from the supervised corpus and unsupervised word representations learned from unannotated corpora. Our models obtain state-of-the-art performance in NER in four languages without resorting to any language-specific knowledge or resources such as gazetteers. 【摘要翻译】最先进的命名实体识别系统在很大程度上依赖于手工创建的特征和特定领域的知识，以便从现有的小型监督训练语料库中有效地学习。在本文中，我们介绍了两种新的神经架构–一种是基于双向 LSTM 和条件随机场的神经架构，另一种是受 shift-reduce 分析器启发，采用基于转换的方法来构建和标记词段的神经架构。我们的模型依赖于两种单词信息来源：从有监督语料库中学到的基于字符的单词表示和从无注释语料库中学到的无监督单词表示。我们的模型在四种语言的 NER 中取得了最先进的性能，而无需借助任何特定语言知识或地名录等资源。
@article{lample2016,
title = {Neural {Architectures} for {Named} {Entity} {Recognition}},
shorttitle = {用于命名实体识别的神经架构},
url = {http://arxiv.org/abs/1603.01360},
abstract = {State-of-the-art named entity recognition systems rely heavily on hand-crafted features and domain-specific knowledge in order to learn effectively from the small, supervised training corpora that are available. In this paper, we introduce two new neural architectures---one based on bidirectional LSTMs and conditional random fields, and the other that constructs and labels segments using a transition-based approach inspired by shift-reduce parsers. Our models rely on two sources of information about words: character-based word representations learned from the supervised corpus and unsupervised word representations learned from unannotated corpora. Our models obtain state-of-the-art performance in NER in four languages without resorting to any language-specific knowledge or resources such as gazetteers.
【摘要翻译】最先进的命名实体识别系统在很大程度上依赖于手工创建的特征和特定领域的知识，以便从现有的小型监督训练语料库中有效地学习。在本文中，我们介绍了两种新的神经架构--一种是基于双向 LSTM 和条件随机场的神经架构，另一种是受 shift-reduce 分析器启发，采用基于转换的方法来构建和标记词段的神经架构。我们的模型依赖于两种单词信息来源：从有监督语料库中学到的基于字符的单词表示和从无注释语料库中学到的无监督单词表示。我们的模型在四种语言的 NER 中取得了最先进的性能，而无需借助任何特定语言知识或地名录等资源。},
language = {en},
urldate = {2018-10-23},
journal = {arXiv:1603.01360 [cs]},
author = {Lample, Guillaume and Ballesteros, Miguel and Subramanian, Sandeep and Kawakami, Kazuya and Dyer, Chris},
month = mar,
year = {2016},
note = {🏷️ /unread、Computer Science - Computation and Language},
keywords = {/unread, Computer Science - Computation and Language},
}
Downloads: 0
{"_id":"y43CPeSnMLXe4SrBD","bibbaseid":"lample-ballesteros-subramanian-kawakami-dyer-neuralarchitecturesfornamedentityrecognition-2016","author_short":["Lample, G.","Ballesteros, M.","Subramanian, S.","Kawakami, K.","Dyer, C."],"bibdata":{"bibtype":"article","type":"article","title":"Neural Architectures for Named Entity Recognition","shorttitle":"用于命名实体识别的神经架构","url":"http://arxiv.org/abs/1603.01360","abstract":"State-of-the-art named entity recognition systems rely heavily on hand-crafted features and domain-specific knowledge in order to learn effectively from the small, supervised training corpora that are available. In this paper, we introduce two new neural architectures—one based on bidirectional LSTMs and conditional random fields, and the other that constructs and labels segments using a transition-based approach inspired by shift-reduce parsers. Our models rely on two sources of information about words: character-based word representations learned from the supervised corpus and unsupervised word representations learned from unannotated corpora. Our models obtain state-of-the-art performance in NER in four languages without resorting to any language-specific knowledge or resources such as gazetteers. 【摘要翻译】最先进的命名实体识别系统在很大程度上依赖于手工创建的特征和特定领域的知识，以便从现有的小型监督训练语料库中有效地学习。在本文中，我们介绍了两种新的神经架构–一种是基于双向 LSTM 和条件随机场的神经架构，另一种是受 shift-reduce 分析器启发，采用基于转换的方法来构建和标记词段的神经架构。我们的模型依赖于两种单词信息来源：从有监督语料库中学到的基于字符的单词表示和从无注释语料库中学到的无监督单词表示。我们的模型在四种语言的 NER 中取得了最先进的性能，而无需借助任何特定语言知识或地名录等资源。","language":"en","urldate":"2018-10-23","journal":"arXiv:1603.01360 [cs]","author":[{"propositions":[],"lastnames":["Lample"],"firstnames":["Guillaume"],"suffixes":[]},{"propositions":[],"lastnames":["Ballesteros"],"firstnames":["Miguel"],"suffixes":[]},{"propositions":[],"lastnames":["Subramanian"],"firstnames":["Sandeep"],"suffixes":[]},{"propositions":[],"lastnames":["Kawakami"],"firstnames":["Kazuya"],"suffixes":[]},{"propositions":[],"lastnames":["Dyer"],"firstnames":["Chris"],"suffixes":[]}],"month":"March","year":"2016","note":"🏷️ /unread、Computer Science - Computation and Language","keywords":"/unread, Computer Science - Computation and Language","bibtex":"@article{lample2016,\n\ttitle = {Neural {Architectures} for {Named} {Entity} {Recognition}},\n\tshorttitle = {用于命名实体识别的神经架构},\n\turl = {http://arxiv.org/abs/1603.01360},\n\tabstract = {State-of-the-art named entity recognition systems rely heavily on hand-crafted features and domain-specific knowledge in order to learn effectively from the small, supervised training corpora that are available. In this paper, we introduce two new neural architectures---one based on bidirectional LSTMs and conditional random fields, and the other that constructs and labels segments using a transition-based approach inspired by shift-reduce parsers. Our models rely on two sources of information about words: character-based word representations learned from the supervised corpus and unsupervised word representations learned from unannotated corpora. Our models obtain state-of-the-art performance in NER in four languages without resorting to any language-specific knowledge or resources such as gazetteers.\n\n【摘要翻译】最先进的命名实体识别系统在很大程度上依赖于手工创建的特征和特定领域的知识，以便从现有的小型监督训练语料库中有效地学习。在本文中，我们介绍了两种新的神经架构--一种是基于双向 LSTM 和条件随机场的神经架构，另一种是受 shift-reduce 分析器启发，采用基于转换的方法来构建和标记词段的神经架构。我们的模型依赖于两种单词信息来源：从有监督语料库中学到的基于字符的单词表示和从无注释语料库中学到的无监督单词表示。我们的模型在四种语言的 NER 中取得了最先进的性能，而无需借助任何特定语言知识或地名录等资源。},\n\tlanguage = {en},\n\turldate = {2018-10-23},\n\tjournal = {arXiv:1603.01360 [cs]},\n\tauthor = {Lample, Guillaume and Ballesteros, Miguel and Subramanian, Sandeep and Kawakami, Kazuya and Dyer, Chris},\n\tmonth = mar,\n\tyear = {2016},\n\tnote = {🏷️ /unread、Computer Science - Computation and Language},\n\tkeywords = {/unread, Computer Science - Computation and Language},\n}\n\n","author_short":["Lample, G.","Ballesteros, M.","Subramanian, S.","Kawakami, K.","Dyer, C."],"key":"lample2016","id":"lample2016","bibbaseid":"lample-ballesteros-subramanian-kawakami-dyer-neuralarchitecturesfornamedentityrecognition-2016","role":"author","urls":{"Paper":"http://arxiv.org/abs/1603.01360"},"keyword":["/unread","Computer Science - Computation and Language"],"metadata":{"authorlinks":{}}},"bibtype":"article","biburl":"https://api.zotero.org/groups/2386895/collections/7PPRTB2H/items?format=bibtex&limit=100","dataSources":["u8q5uny4m5jJL9RcX"],"keywords":["/unread","computer science - computation and language"],"search_terms":["neural","architectures","named","entity","recognition","lample","ballesteros","subramanian","kawakami","dyer"],"title":"Neural Architectures for Named Entity Recognition","year":2016}