@inproceedings{laicher-etal-2021-explaining,
title = "Explaining and Improving {BERT} Performance on Lexical Semantic Change Detection",
author = "Laicher, Severin and
Kurtyigit, Sinan and
Schlechtweg, Dominik and
Kuhn, Jonas and
Schulte im Walde, Sabine",
editor = "Sorodoc, Ionut-Teodor and
Sushil, Madhumita and
Takmaz, Ece and
Agirre, Eneko",
booktitle = "Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Student Research Workshop",
month = apr,
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2021.eacl-srw.25",
doi = "10.18653/v1/2021.eacl-srw.25",
pages = "192--202",
abstract = "Type- and token-based embedding architectures are still competing in lexical semantic change detection. The recent success of type-based models in SemEval-2020 Task 1 has raised the question why the success of token-based models on a variety of other NLP tasks does not translate to our field. We investigate the influence of a range of variables on clusterings of BERT vectors and show that its low performance is largely due to orthographic information on the target word, which is encoded even in the higher layers of BERT representations. By reducing the influence of orthography we considerably improve BERT{'}s performance.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="laicher-etal-2021-explaining">
<titleInfo>
<title>Explaining and Improving BERT Performance on Lexical Semantic Change Detection</title>
</titleInfo>
<name type="personal">
<namePart type="given">Severin</namePart>
<namePart type="family">Laicher</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sinan</namePart>
<namePart type="family">Kurtyigit</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Dominik</namePart>
<namePart type="family">Schlechtweg</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jonas</namePart>
<namePart type="family">Kuhn</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sabine</namePart>
<namePart type="family">Schulte im Walde</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2021-04</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Student Research Workshop</title>
</titleInfo>
<name type="personal">
<namePart type="given">Ionut-Teodor</namePart>
<namePart type="family">Sorodoc</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Madhumita</namePart>
<namePart type="family">Sushil</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ece</namePart>
<namePart type="family">Takmaz</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Eneko</namePart>
<namePart type="family">Agirre</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Online</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Type- and token-based embedding architectures are still competing in lexical semantic change detection. The recent success of type-based models in SemEval-2020 Task 1 has raised the question why the success of token-based models on a variety of other NLP tasks does not translate to our field. We investigate the influence of a range of variables on clusterings of BERT vectors and show that its low performance is largely due to orthographic information on the target word, which is encoded even in the higher layers of BERT representations. By reducing the influence of orthography we considerably improve BERT’s performance.</abstract>
<identifier type="citekey">laicher-etal-2021-explaining</identifier>
<identifier type="doi">10.18653/v1/2021.eacl-srw.25</identifier>
<location>
<url>https://aclanthology.org/2021.eacl-srw.25</url>
</location>
<part>
<date>2021-04</date>
<extent unit="page">
<start>192</start>
<end>202</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Explaining and Improving BERT Performance on Lexical Semantic Change Detection
%A Laicher, Severin
%A Kurtyigit, Sinan
%A Schlechtweg, Dominik
%A Kuhn, Jonas
%A Schulte im Walde, Sabine
%Y Sorodoc, Ionut-Teodor
%Y Sushil, Madhumita
%Y Takmaz, Ece
%Y Agirre, Eneko
%S Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Student Research Workshop
%D 2021
%8 April
%I Association for Computational Linguistics
%C Online
%F laicher-etal-2021-explaining
%X Type- and token-based embedding architectures are still competing in lexical semantic change detection. The recent success of type-based models in SemEval-2020 Task 1 has raised the question why the success of token-based models on a variety of other NLP tasks does not translate to our field. We investigate the influence of a range of variables on clusterings of BERT vectors and show that its low performance is largely due to orthographic information on the target word, which is encoded even in the higher layers of BERT representations. By reducing the influence of orthography we considerably improve BERT’s performance.
%R 10.18653/v1/2021.eacl-srw.25
%U https://aclanthology.org/2021.eacl-srw.25
%U https://doi.org/10.18653/v1/2021.eacl-srw.25
%P 192-202
Markdown (Informal)
[Explaining and Improving BERT Performance on Lexical Semantic Change Detection](https://aclanthology.org/2021.eacl-srw.25) (Laicher et al., EACL 2021)
ACL