@inproceedings{kutuzov-kuzmenko-2019-lemmatize,
title = "To Lemmatize or Not to Lemmatize: How Word Normalisation Affects {ELM}o Performance in Word Sense Disambiguation",
author = "Kutuzov, Andrey and
Kuzmenko, Elizaveta",
editor = {Nivre, Joakim and
Derczynski, Leon and
Ginter, Filip and
Lindi, Bj{\o}rn and
Oepen, Stephan and
S{\o}gaard, Anders and
Tidemann, J{\"o}rg},
booktitle = "Proceedings of the First NLPL Workshop on Deep Learning for Natural Language Processing",
month = sep,
year = "2019",
address = "Turku, Finland",
publisher = {Link{\"o}ping University Electronic Press},
url = "https://aclanthology.org/W19-6203",
pages = "22--28",
abstract = "In this paper, we critically evaluate the widespread assumption that deep learning NLP models do not require lemmatized input. To test this, we trained versions of contextualised word embedding ELMo models on raw tokenized corpora and on the corpora with word tokens replaced by their lemmas. Then, these models were evaluated on the word sense disambiguation task. This was done for the English and Russian languages. The experiments showed that while lemmatization is indeed not necessary for English, the situation is different for Russian. It seems that for rich-morphology languages, using lemmatized training and testing data yields small but consistent improvements: at least for word sense disambiguation. This means that the decisions about text pre-processing before training ELMo should consider the linguistic nature of the language in question.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="kutuzov-kuzmenko-2019-lemmatize">
<titleInfo>
<title>To Lemmatize or Not to Lemmatize: How Word Normalisation Affects ELMo Performance in Word Sense Disambiguation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Andrey</namePart>
<namePart type="family">Kutuzov</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Elizaveta</namePart>
<namePart type="family">Kuzmenko</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2019-09</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the First NLPL Workshop on Deep Learning for Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Joakim</namePart>
<namePart type="family">Nivre</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Leon</namePart>
<namePart type="family">Derczynski</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Filip</namePart>
<namePart type="family">Ginter</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Bjørn</namePart>
<namePart type="family">Lindi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Stephan</namePart>
<namePart type="family">Oepen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Anders</namePart>
<namePart type="family">Søgaard</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jörg</namePart>
<namePart type="family">Tidemann</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Linköping University Electronic Press</publisher>
<place>
<placeTerm type="text">Turku, Finland</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>In this paper, we critically evaluate the widespread assumption that deep learning NLP models do not require lemmatized input. To test this, we trained versions of contextualised word embedding ELMo models on raw tokenized corpora and on the corpora with word tokens replaced by their lemmas. Then, these models were evaluated on the word sense disambiguation task. This was done for the English and Russian languages. The experiments showed that while lemmatization is indeed not necessary for English, the situation is different for Russian. It seems that for rich-morphology languages, using lemmatized training and testing data yields small but consistent improvements: at least for word sense disambiguation. This means that the decisions about text pre-processing before training ELMo should consider the linguistic nature of the language in question.</abstract>
<identifier type="citekey">kutuzov-kuzmenko-2019-lemmatize</identifier>
<location>
<url>https://aclanthology.org/W19-6203</url>
</location>
<part>
<date>2019-09</date>
<extent unit="page">
<start>22</start>
<end>28</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T To Lemmatize or Not to Lemmatize: How Word Normalisation Affects ELMo Performance in Word Sense Disambiguation
%A Kutuzov, Andrey
%A Kuzmenko, Elizaveta
%Y Nivre, Joakim
%Y Derczynski, Leon
%Y Ginter, Filip
%Y Lindi, Bjørn
%Y Oepen, Stephan
%Y Søgaard, Anders
%Y Tidemann, Jörg
%S Proceedings of the First NLPL Workshop on Deep Learning for Natural Language Processing
%D 2019
%8 September
%I Linköping University Electronic Press
%C Turku, Finland
%F kutuzov-kuzmenko-2019-lemmatize
%X In this paper, we critically evaluate the widespread assumption that deep learning NLP models do not require lemmatized input. To test this, we trained versions of contextualised word embedding ELMo models on raw tokenized corpora and on the corpora with word tokens replaced by their lemmas. Then, these models were evaluated on the word sense disambiguation task. This was done for the English and Russian languages. The experiments showed that while lemmatization is indeed not necessary for English, the situation is different for Russian. It seems that for rich-morphology languages, using lemmatized training and testing data yields small but consistent improvements: at least for word sense disambiguation. This means that the decisions about text pre-processing before training ELMo should consider the linguistic nature of the language in question.
%U https://aclanthology.org/W19-6203
%P 22-28
Markdown (Informal)
[To Lemmatize or Not to Lemmatize: How Word Normalisation Affects ELMo Performance in Word Sense Disambiguation](https://aclanthology.org/W19-6203) (Kutuzov & Kuzmenko, NoDaLiDa 2019)
ACL