@inproceedings{wu-etal-2022-fine,
title = "Fine-grained Multi-lingual Disentangled Autoencoder for Language-agnostic Representation Learning",
author = "Wu, Zetian and
Sun, Zhongkai and
Zhao, Zhengyang and
Lu, Sixing and
Ma, Chengyuan and
Guo, Chenlei",
editor = "FitzGerald, Jack and
Rottmann, Kay and
Hirschberg, Julia and
Bansal, Mohit and
Rumshisky, Anna and
Peris, Charith and
Hench, Christopher",
booktitle = "Proceedings of the Massively Multilingual Natural Language Understanding Workshop (MMNLU-22)",
month = dec,
year = "2022",
address = "Abu Dhabi, United Arab Emirates (Hybrid)",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2022.mmnlu-1.2",
doi = "10.18653/v1/2022.mmnlu-1.2",
pages = "12--24",
abstract = "Encoding both language-specific and language-agnostic information into a single high-dimensional space is a common practice of pre-trained Multi-lingual Language Models (pMLM). Such encoding has been shown to perform effectively on natural language tasks requiring semantics of the whole sentence (e.g., translation). However, its effectiveness appears to be limited on tasks requiring partial information of the utterance (e.g., multi-lingual entity retrieval, template retrieval, and semantic alignment). In this work, a novel Fine-grained Multilingual Disentangled Autoencoder (FMDA) is proposed to disentangle fine-grained semantic information from language-specific information in a multi-lingual setting. FMDA is capable of successfully extracting the disentangled template semantic and residual semantic representations. Experiments conducted on the MASSIVE dataset demonstrate that the disentangled encoding can boost each other during the training, thus consistently outperforming the original pMLM and the strong language disentanglement baseline on monolingual template retrieval and cross-lingual semantic retrieval tasks across multiple languages.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="wu-etal-2022-fine">
<titleInfo>
<title>Fine-grained Multi-lingual Disentangled Autoencoder for Language-agnostic Representation Learning</title>
</titleInfo>
<name type="personal">
<namePart type="given">Zetian</namePart>
<namePart type="family">Wu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhongkai</namePart>
<namePart type="family">Sun</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhengyang</namePart>
<namePart type="family">Zhao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sixing</namePart>
<namePart type="family">Lu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chengyuan</namePart>
<namePart type="family">Ma</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chenlei</namePart>
<namePart type="family">Guo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2022-12</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the Massively Multilingual Natural Language Understanding Workshop (MMNLU-22)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Jack</namePart>
<namePart type="family">FitzGerald</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kay</namePart>
<namePart type="family">Rottmann</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Julia</namePart>
<namePart type="family">Hirschberg</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mohit</namePart>
<namePart type="family">Bansal</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Anna</namePart>
<namePart type="family">Rumshisky</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Charith</namePart>
<namePart type="family">Peris</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Christopher</namePart>
<namePart type="family">Hench</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Abu Dhabi, United Arab Emirates (Hybrid)</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Encoding both language-specific and language-agnostic information into a single high-dimensional space is a common practice of pre-trained Multi-lingual Language Models (pMLM). Such encoding has been shown to perform effectively on natural language tasks requiring semantics of the whole sentence (e.g., translation). However, its effectiveness appears to be limited on tasks requiring partial information of the utterance (e.g., multi-lingual entity retrieval, template retrieval, and semantic alignment). In this work, a novel Fine-grained Multilingual Disentangled Autoencoder (FMDA) is proposed to disentangle fine-grained semantic information from language-specific information in a multi-lingual setting. FMDA is capable of successfully extracting the disentangled template semantic and residual semantic representations. Experiments conducted on the MASSIVE dataset demonstrate that the disentangled encoding can boost each other during the training, thus consistently outperforming the original pMLM and the strong language disentanglement baseline on monolingual template retrieval and cross-lingual semantic retrieval tasks across multiple languages.</abstract>
<identifier type="citekey">wu-etal-2022-fine</identifier>
<identifier type="doi">10.18653/v1/2022.mmnlu-1.2</identifier>
<location>
<url>https://aclanthology.org/2022.mmnlu-1.2</url>
</location>
<part>
<date>2022-12</date>
<extent unit="page">
<start>12</start>
<end>24</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Fine-grained Multi-lingual Disentangled Autoencoder for Language-agnostic Representation Learning
%A Wu, Zetian
%A Sun, Zhongkai
%A Zhao, Zhengyang
%A Lu, Sixing
%A Ma, Chengyuan
%A Guo, Chenlei
%Y FitzGerald, Jack
%Y Rottmann, Kay
%Y Hirschberg, Julia
%Y Bansal, Mohit
%Y Rumshisky, Anna
%Y Peris, Charith
%Y Hench, Christopher
%S Proceedings of the Massively Multilingual Natural Language Understanding Workshop (MMNLU-22)
%D 2022
%8 December
%I Association for Computational Linguistics
%C Abu Dhabi, United Arab Emirates (Hybrid)
%F wu-etal-2022-fine
%X Encoding both language-specific and language-agnostic information into a single high-dimensional space is a common practice of pre-trained Multi-lingual Language Models (pMLM). Such encoding has been shown to perform effectively on natural language tasks requiring semantics of the whole sentence (e.g., translation). However, its effectiveness appears to be limited on tasks requiring partial information of the utterance (e.g., multi-lingual entity retrieval, template retrieval, and semantic alignment). In this work, a novel Fine-grained Multilingual Disentangled Autoencoder (FMDA) is proposed to disentangle fine-grained semantic information from language-specific information in a multi-lingual setting. FMDA is capable of successfully extracting the disentangled template semantic and residual semantic representations. Experiments conducted on the MASSIVE dataset demonstrate that the disentangled encoding can boost each other during the training, thus consistently outperforming the original pMLM and the strong language disentanglement baseline on monolingual template retrieval and cross-lingual semantic retrieval tasks across multiple languages.
%R 10.18653/v1/2022.mmnlu-1.2
%U https://aclanthology.org/2022.mmnlu-1.2
%U https://doi.org/10.18653/v1/2022.mmnlu-1.2
%P 12-24
Markdown (Informal)
[Fine-grained Multi-lingual Disentangled Autoencoder for Language-agnostic Representation Learning](https://aclanthology.org/2022.mmnlu-1.2) (Wu et al., MMNLU 2022)
ACL