@inproceedings{rachinskiy-arefyev-2022-black,
title = "{G}loss{R}eader at {LSCD}iscovery: Train to Select a Proper Gloss in {E}nglish {--} Discover Lexical Semantic Change in {S}panish",
author = "Rachinskiy, Maxim and
Arefyev, Nikolay",
editor = "Tahmasebi, Nina and
Montariol, Syrielle and
Kutuzov, Andrey and
Hengchen, Simon and
Dubossarsky, Haim and
Borin, Lars",
booktitle = "Proceedings of the 3rd Workshop on Computational Approaches to Historical Language Change",
month = may,
year = "2022",
address = "Dublin, Ireland",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2022.lchange-1.22",
doi = "10.18653/v1/2022.lchange-1.22",
pages = "198--203",
abstract = "The contextualized embeddings obtained from neural networks pre-trained as Language Models (LM) or Masked Language Models (MLM) are not well suitable for solving the Lexical Semantic Change Detection (LSCD) task because they are more sensitive to changes in word forms rather than word meaning, a property previously known as the word form bias or orthographic bias. Unlike many other NLP tasks, it is also not obvious how to fine-tune such models for LSCD. In order to conclude if there are any differences between senses of a particular word in two corpora, a human annotator or a system shall analyze many examples containing this word from both corpora. This makes annotation of LSCD datasets very labour-consuming. The existing LSCD datasets contain up to 100 words that are labeled according to their semantic change, which is hardly enough for fine-tuning. To solve these problems we fine-tune the XLM-R MLM as part of a gloss-based WSD system on a large WSD dataset in English. Then we employ zero-shot cross-lingual transferability of XLM-R to build the contextualized embeddings for examples in Spanish. In order to obtain the graded change score for each word, we calculate the average distance between our improved contextualized embeddings of its old and new occurrences. For the binary change detection subtask, we apply thresholding to the same scores. Our solution has shown the best results among all other participants in all subtasks except for the optional sense gain detection subtask.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="rachinskiy-arefyev-2022-black">
<titleInfo>
<title>GlossReader at LSCDiscovery: Train to Select a Proper Gloss in English – Discover Lexical Semantic Change in Spanish</title>
</titleInfo>
<name type="personal">
<namePart type="given">Maxim</namePart>
<namePart type="family">Rachinskiy</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nikolay</namePart>
<namePart type="family">Arefyev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2022-05</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 3rd Workshop on Computational Approaches to Historical Language Change</title>
</titleInfo>
<name type="personal">
<namePart type="given">Nina</namePart>
<namePart type="family">Tahmasebi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Syrielle</namePart>
<namePart type="family">Montariol</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Andrey</namePart>
<namePart type="family">Kutuzov</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Simon</namePart>
<namePart type="family">Hengchen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Haim</namePart>
<namePart type="family">Dubossarsky</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lars</namePart>
<namePart type="family">Borin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Dublin, Ireland</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>The contextualized embeddings obtained from neural networks pre-trained as Language Models (LM) or Masked Language Models (MLM) are not well suitable for solving the Lexical Semantic Change Detection (LSCD) task because they are more sensitive to changes in word forms rather than word meaning, a property previously known as the word form bias or orthographic bias. Unlike many other NLP tasks, it is also not obvious how to fine-tune such models for LSCD. In order to conclude if there are any differences between senses of a particular word in two corpora, a human annotator or a system shall analyze many examples containing this word from both corpora. This makes annotation of LSCD datasets very labour-consuming. The existing LSCD datasets contain up to 100 words that are labeled according to their semantic change, which is hardly enough for fine-tuning. To solve these problems we fine-tune the XLM-R MLM as part of a gloss-based WSD system on a large WSD dataset in English. Then we employ zero-shot cross-lingual transferability of XLM-R to build the contextualized embeddings for examples in Spanish. In order to obtain the graded change score for each word, we calculate the average distance between our improved contextualized embeddings of its old and new occurrences. For the binary change detection subtask, we apply thresholding to the same scores. Our solution has shown the best results among all other participants in all subtasks except for the optional sense gain detection subtask.</abstract>
<identifier type="citekey">rachinskiy-arefyev-2022-black</identifier>
<identifier type="doi">10.18653/v1/2022.lchange-1.22</identifier>
<location>
<url>https://aclanthology.org/2022.lchange-1.22</url>
</location>
<part>
<date>2022-05</date>
<extent unit="page">
<start>198</start>
<end>203</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T GlossReader at LSCDiscovery: Train to Select a Proper Gloss in English – Discover Lexical Semantic Change in Spanish
%A Rachinskiy, Maxim
%A Arefyev, Nikolay
%Y Tahmasebi, Nina
%Y Montariol, Syrielle
%Y Kutuzov, Andrey
%Y Hengchen, Simon
%Y Dubossarsky, Haim
%Y Borin, Lars
%S Proceedings of the 3rd Workshop on Computational Approaches to Historical Language Change
%D 2022
%8 May
%I Association for Computational Linguistics
%C Dublin, Ireland
%F rachinskiy-arefyev-2022-black
%X The contextualized embeddings obtained from neural networks pre-trained as Language Models (LM) or Masked Language Models (MLM) are not well suitable for solving the Lexical Semantic Change Detection (LSCD) task because they are more sensitive to changes in word forms rather than word meaning, a property previously known as the word form bias or orthographic bias. Unlike many other NLP tasks, it is also not obvious how to fine-tune such models for LSCD. In order to conclude if there are any differences between senses of a particular word in two corpora, a human annotator or a system shall analyze many examples containing this word from both corpora. This makes annotation of LSCD datasets very labour-consuming. The existing LSCD datasets contain up to 100 words that are labeled according to their semantic change, which is hardly enough for fine-tuning. To solve these problems we fine-tune the XLM-R MLM as part of a gloss-based WSD system on a large WSD dataset in English. Then we employ zero-shot cross-lingual transferability of XLM-R to build the contextualized embeddings for examples in Spanish. In order to obtain the graded change score for each word, we calculate the average distance between our improved contextualized embeddings of its old and new occurrences. For the binary change detection subtask, we apply thresholding to the same scores. Our solution has shown the best results among all other participants in all subtasks except for the optional sense gain detection subtask.
%R 10.18653/v1/2022.lchange-1.22
%U https://aclanthology.org/2022.lchange-1.22
%U https://doi.org/10.18653/v1/2022.lchange-1.22
%P 198-203
Markdown (Informal)
[GlossReader at LSCDiscovery: Train to Select a Proper Gloss in English – Discover Lexical Semantic Change in Spanish](https://aclanthology.org/2022.lchange-1.22) (Rachinskiy & Arefyev, LChange 2022)
ACL