@inproceedings{acikgoz-etal-2022-transformers,
title = "Transformers on Multilingual Clause-Level Morphology",
author = {Acikgoz, Emre Can and
Chubakov, Tilek and
Kural, Muge and
{\c{S}}ahin, G{\"o}zde and
Yuret, Deniz},
editor = {Ataman, Duygu and
Gonen, Hila and
Ruder, Sebastian and
Firat, Orhan and
G{\"u}l Sahin, G{\"o}zde and
Mirzakhalov, Jamshidbek},
booktitle = "Proceedings of the 2nd Workshop on Multi-lingual Representation Learning (MRL)",
month = dec,
year = "2022",
address = "Abu Dhabi, United Arab Emirates (Hybrid)",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2022.mrl-1.10/",
doi = "10.18653/v1/2022.mrl-1.10",
pages = "100--105",
abstract = "This paper describes the KUIS-AI NLP team`s submission for the 1st Shared Task on Multilingual Clause-level Morphology (MRL2022). We present our work on all three parts of the shared task: inflection, reinflection, and analysis. We mainly explore two approaches: Trans- former models in combination with data augmentation, and exploiting the state-of-the-art language modeling techniques for morphological analysis. Data augmentation leads to a remarkable performance improvement for most of the languages in the inflection task. Prefix-tuning on pretrained mGPT model helps us to adapt reinflection and analysis tasks in a low-data setting. Additionally, we used pipeline architectures using publicly available open-source lemmatization tools and monolingual BERT- based morphological feature classifiers for rein- flection and analysis tasks, respectively. While Transformer architectures with data augmentation and pipeline architectures achieved the best results for inflection and reinflection tasks, pipelines and prefix-tuning on mGPT received the highest results for the analysis task. Our methods achieved first place in each of the three tasks and outperforms mT5-baseline with 89{\%} for inflection, 80{\%} for reflection, and 12{\%} for analysis. Our code 1 is publicly available."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="acikgoz-etal-2022-transformers">
<titleInfo>
<title>Transformers on Multilingual Clause-Level Morphology</title>
</titleInfo>
<name type="personal">
<namePart type="given">Emre</namePart>
<namePart type="given">Can</namePart>
<namePart type="family">Acikgoz</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tilek</namePart>
<namePart type="family">Chubakov</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Muge</namePart>
<namePart type="family">Kural</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Gözde</namePart>
<namePart type="family">Şahin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Deniz</namePart>
<namePart type="family">Yuret</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 2nd Workshop on Multi-lingual Representation Learning (MRL)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Duygu</namePart>
<namePart type="family">Ataman</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hila</namePart>
<namePart type="family">Gonen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sebastian</namePart>
<namePart type="family">Ruder</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Orhan</namePart>
<namePart type="family">Firat</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Gözde</namePart>
<namePart type="family">Gül Sahin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jamshidbek</namePart>
<namePart type="family">Mirzakhalov</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>This paper describes the KUIS-AI NLP team‘s submission for the 1st Shared Task on Multilingual Clause-level Morphology (MRL2022). We present our work on all three parts of the shared task: inflection, reinflection, and analysis. We mainly explore two approaches: Trans- former models in combination with data augmentation, and exploiting the state-of-the-art language modeling techniques for morphological analysis. Data augmentation leads to a remarkable performance improvement for most of the languages in the inflection task. Prefix-tuning on pretrained mGPT model helps us to adapt reinflection and analysis tasks in a low-data setting. Additionally, we used pipeline architectures using publicly available open-source lemmatization tools and monolingual BERT- based morphological feature classifiers for rein- flection and analysis tasks, respectively. While Transformer architectures with data augmentation and pipeline architectures achieved the best results for inflection and reinflection tasks, pipelines and prefix-tuning on mGPT received the highest results for the analysis task. Our methods achieved first place in each of the three tasks and outperforms mT5-baseline with 89% for inflection, 80% for reflection, and 12% for analysis. Our code 1 is publicly available.</abstract>
<identifier type="citekey">acikgoz-etal-2022-transformers</identifier>
<identifier type="doi">10.18653/v1/2022.mrl-1.10</identifier>
<location>
<url>https://aclanthology.org/2022.mrl-1.10/</url>
</location>
<part>
<date>2022-12</date>
<extent unit="page">
<start>100</start>
<end>105</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Transformers on Multilingual Clause-Level Morphology
%A Acikgoz, Emre Can
%A Chubakov, Tilek
%A Kural, Muge
%A Şahin, Gözde
%A Yuret, Deniz
%Y Ataman, Duygu
%Y Gonen, Hila
%Y Ruder, Sebastian
%Y Firat, Orhan
%Y Gül Sahin, Gözde
%Y Mirzakhalov, Jamshidbek
%S Proceedings of the 2nd Workshop on Multi-lingual Representation Learning (MRL)
%D 2022
%8 December
%I Association for Computational Linguistics
%C Abu Dhabi, United Arab Emirates (Hybrid)
%F acikgoz-etal-2022-transformers
%X This paper describes the KUIS-AI NLP team‘s submission for the 1st Shared Task on Multilingual Clause-level Morphology (MRL2022). We present our work on all three parts of the shared task: inflection, reinflection, and analysis. We mainly explore two approaches: Trans- former models in combination with data augmentation, and exploiting the state-of-the-art language modeling techniques for morphological analysis. Data augmentation leads to a remarkable performance improvement for most of the languages in the inflection task. Prefix-tuning on pretrained mGPT model helps us to adapt reinflection and analysis tasks in a low-data setting. Additionally, we used pipeline architectures using publicly available open-source lemmatization tools and monolingual BERT- based morphological feature classifiers for rein- flection and analysis tasks, respectively. While Transformer architectures with data augmentation and pipeline architectures achieved the best results for inflection and reinflection tasks, pipelines and prefix-tuning on mGPT received the highest results for the analysis task. Our methods achieved first place in each of the three tasks and outperforms mT5-baseline with 89% for inflection, 80% for reflection, and 12% for analysis. Our code 1 is publicly available.
%R 10.18653/v1/2022.mrl-1.10
%U https://aclanthology.org/2022.mrl-1.10/
%U https://doi.org/10.18653/v1/2022.mrl-1.10
%P 100-105
Markdown (Informal)
[Transformers on Multilingual Clause-Level Morphology](https://aclanthology.org/2022.mrl-1.10/) (Acikgoz et al., MRL 2022)
ACL
- Emre Can Acikgoz, Tilek Chubakov, Muge Kural, Gözde Şahin, and Deniz Yuret. 2022. Transformers on Multilingual Clause-Level Morphology. In Proceedings of the 2nd Workshop on Multi-lingual Representation Learning (MRL), pages 100–105, Abu Dhabi, United Arab Emirates (Hybrid). Association for Computational Linguistics.