Deep contextualized word representations. Peters, M. E., Neumann, M., Iyyer, M., Gardner, M., Clark, C., Lee, K., & Zettlemoyer, L. In Proceedings of North American Association for Computational Linguistics (NAACL), feb, 2018. ![Deep contextualized word representations [link]](https://bibbase.org/img/filetypes/link.svg "link")
Paper abstract bibtex We introduce a new type of deep contextualized word representation that models both (1) complex characteristics of word use (e.g., syntax and semantics), and (2) how these uses vary across linguistic contexts (i.e., to model polysemy). Our word vectors are learned functions of the internal states of a deep bidirectional language model (biLM), which is pre-trained on a large text corpus. We show that these representations can be easily added to existing models and significantly improve the state of the art across six challenging NLP problems, including question answering, textual entailment and sentiment analysis. We also present an analysis showing that exposing the deep internals of the pre-trained network is crucial, allowing downstream models to mix different types of semi-supervision signals.
@inproceedings{Peters2018,
abstract = {We introduce a new type of deep contextualized word representation that models both (1) complex characteristics of word use (e.g., syntax and semantics), and (2) how these uses vary across linguistic contexts (i.e., to model polysemy). Our word vectors are learned functions of the internal states of a deep bidirectional language model (biLM), which is pre-trained on a large text corpus. We show that these representations can be easily added to existing models and significantly improve the state of the art across six challenging NLP problems, including question answering, textual entailment and sentiment analysis. We also present an analysis showing that exposing the deep internals of the pre-trained network is crucial, allowing downstream models to mix different types of semi-supervision signals.},
archivePrefix = {arXiv},
arxivId = {1802.05365},
author = {Peters, Matthew E. and Neumann, Mark and Iyyer, Mohit and Gardner, Matt and Clark, Christopher and Lee, Kenton and Zettlemoyer, Luke},
booktitle = {Proceedings of North American Association for Computational Linguistics (NAACL)},
eprint = {1802.05365},
file = {:Users/shanest/Documents/Library/Peters et al/Proceedings of North American Association for Computational Linguistics (NAACL)/Peters et al. - 2018 - Deep contextualized word representations.pdf:pdf},
keywords = {model},
month = {feb},
title = {{Deep contextualized word representations}},
url = {http://arxiv.org/abs/1802.05365},
year = {2018}
}
Downloads: 0
{"_id":"BpsqGRny5raSWf6cC","bibbaseid":"peters-neumann-iyyer-gardner-clark-lee-zettlemoyer-deepcontextualizedwordrepresentations-2018","authorIDs":[],"author_short":["Peters, M. E.","Neumann, M.","Iyyer, M.","Gardner, M.","Clark, C.","Lee, K.","Zettlemoyer, L."],"bibdata":{"bibtype":"inproceedings","type":"inproceedings","abstract":"We introduce a new type of deep contextualized word representation that models both (1) complex characteristics of word use (e.g., syntax and semantics), and (2) how these uses vary across linguistic contexts (i.e., to model polysemy). Our word vectors are learned functions of the internal states of a deep bidirectional language model (biLM), which is pre-trained on a large text corpus. We show that these representations can be easily added to existing models and significantly improve the state of the art across six challenging NLP problems, including question answering, textual entailment and sentiment analysis. We also present an analysis showing that exposing the deep internals of the pre-trained network is crucial, allowing downstream models to mix different types of semi-supervision signals.","archiveprefix":"arXiv","arxivid":"1802.05365","author":[{"propositions":[],"lastnames":["Peters"],"firstnames":["Matthew","E."],"suffixes":[]},{"propositions":[],"lastnames":["Neumann"],"firstnames":["Mark"],"suffixes":[]},{"propositions":[],"lastnames":["Iyyer"],"firstnames":["Mohit"],"suffixes":[]},{"propositions":[],"lastnames":["Gardner"],"firstnames":["Matt"],"suffixes":[]},{"propositions":[],"lastnames":["Clark"],"firstnames":["Christopher"],"suffixes":[]},{"propositions":[],"lastnames":["Lee"],"firstnames":["Kenton"],"suffixes":[]},{"propositions":[],"lastnames":["Zettlemoyer"],"firstnames":["Luke"],"suffixes":[]}],"booktitle":"Proceedings of North American Association for Computational Linguistics (NAACL)","eprint":"1802.05365","file":":Users/shanest/Documents/Library/Peters et al/Proceedings of North American Association for Computational Linguistics (NAACL)/Peters et al. - 2018 - Deep contextualized word representations.pdf:pdf","keywords":"model","month":"feb","title":"Deep contextualized word representations","url":"http://arxiv.org/abs/1802.05365","year":"2018","bibtex":"@inproceedings{Peters2018,\nabstract = {We introduce a new type of deep contextualized word representation that models both (1) complex characteristics of word use (e.g., syntax and semantics), and (2) how these uses vary across linguistic contexts (i.e., to model polysemy). Our word vectors are learned functions of the internal states of a deep bidirectional language model (biLM), which is pre-trained on a large text corpus. We show that these representations can be easily added to existing models and significantly improve the state of the art across six challenging NLP problems, including question answering, textual entailment and sentiment analysis. We also present an analysis showing that exposing the deep internals of the pre-trained network is crucial, allowing downstream models to mix different types of semi-supervision signals.},\narchivePrefix = {arXiv},\narxivId = {1802.05365},\nauthor = {Peters, Matthew E. and Neumann, Mark and Iyyer, Mohit and Gardner, Matt and Clark, Christopher and Lee, Kenton and Zettlemoyer, Luke},\nbooktitle = {Proceedings of North American Association for Computational Linguistics (NAACL)},\neprint = {1802.05365},\nfile = {:Users/shanest/Documents/Library/Peters et al/Proceedings of North American Association for Computational Linguistics (NAACL)/Peters et al. - 2018 - Deep contextualized word representations.pdf:pdf},\nkeywords = {model},\nmonth = {feb},\ntitle = {{Deep contextualized word representations}},\nurl = {http://arxiv.org/abs/1802.05365},\nyear = {2018}\n}\n","author_short":["Peters, M. E.","Neumann, M.","Iyyer, M.","Gardner, M.","Clark, C.","Lee, K.","Zettlemoyer, L."],"key":"Peters2018","id":"Peters2018","bibbaseid":"peters-neumann-iyyer-gardner-clark-lee-zettlemoyer-deepcontextualizedwordrepresentations-2018","role":"author","urls":{"Paper":"http://arxiv.org/abs/1802.05365"},"keyword":["model"],"metadata":{"authorlinks":{}},"downloads":0},"bibtype":"inproceedings","biburl":"https://www.shane.st/teaching/575/win20/MachineLearning-interpretability.bib","creationDate":"2020-01-05T04:04:02.876Z","downloads":0,"keywords":["model"],"search_terms":["deep","contextualized","word","representations","peters","neumann","iyyer","gardner","clark","lee","zettlemoyer"],"title":"Deep contextualized word representations","year":2018,"dataSources":["okYcdTpf4JJ2zkj7A","QGwcHf7xnb5mCCQi7","FoqTuDx9Cxduvwxiv","aXmRAq63YsH7a3ufx","znj7izS5PeehdLR3G"]}