@inproceedings{amestica-etal-2025-joint,
title = "A Joint Multitask Model for Morpho-Syntactic Parsing",
author = "Am{\'e}stica, Demian Inostroza and
Mistica, Meladel and
Vylomova, Ekaterina and
Guest, Chris and
Kurniawan, Kemal",
editor = "Goldman, Omer and
Weissweiler, Leonie and
Tsarfaty, Reut",
booktitle = "Proceedings of The UniDive 2025 Shared Task on Multilingual Morpho-Syntactic Parsing",
month = aug,
year = "2025",
address = "Ljubljana, Slovenia",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.unidive-1.2/",
pages = "19--26",
ISBN = "979-8-89176-320-3",
abstract = "We present a joint multitask model for the Uni-Dive 2025 Morpho-Syntactic Parsing shared task, where systems predict both morphological and syntactic analyses following novel UD annotation scheme. Our system uses a shared XLM-RoBERTa encoder with three specialized decoders for content word identification, dependency parsing, and morphosyntactic feature prediction. Our model achieves the best overall performance on the shared task{'}s leaderboard covering nine typologically diverse languages, with an average MSLAS score of 78.7{\%}, LAS of 80.1{\%}, and Feats F1 of 90.3{\%}. Our ablation studies show that matching the task{'}s gold tokenization and content word identification are crucial to model performance. Error analysis reveals that our model struggles with core grammatical cases (particularly Nom{--}Acc) and nominal features across languages."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="amestica-etal-2025-joint">
<titleInfo>
<title>A Joint Multitask Model for Morpho-Syntactic Parsing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Demian</namePart>
<namePart type="given">Inostroza</namePart>
<namePart type="family">Améstica</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Meladel</namePart>
<namePart type="family">Mistica</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ekaterina</namePart>
<namePart type="family">Vylomova</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chris</namePart>
<namePart type="family">Guest</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kemal</namePart>
<namePart type="family">Kurniawan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-08</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of The UniDive 2025 Shared Task on Multilingual Morpho-Syntactic Parsing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Omer</namePart>
<namePart type="family">Goldman</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Leonie</namePart>
<namePart type="family">Weissweiler</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Reut</namePart>
<namePart type="family">Tsarfaty</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Ljubljana, Slovenia</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-320-3</identifier>
</relatedItem>
<abstract>We present a joint multitask model for the Uni-Dive 2025 Morpho-Syntactic Parsing shared task, where systems predict both morphological and syntactic analyses following novel UD annotation scheme. Our system uses a shared XLM-RoBERTa encoder with three specialized decoders for content word identification, dependency parsing, and morphosyntactic feature prediction. Our model achieves the best overall performance on the shared task’s leaderboard covering nine typologically diverse languages, with an average MSLAS score of 78.7%, LAS of 80.1%, and Feats F1 of 90.3%. Our ablation studies show that matching the task’s gold tokenization and content word identification are crucial to model performance. Error analysis reveals that our model struggles with core grammatical cases (particularly Nom–Acc) and nominal features across languages.</abstract>
<identifier type="citekey">amestica-etal-2025-joint</identifier>
<location>
<url>https://aclanthology.org/2025.unidive-1.2/</url>
</location>
<part>
<date>2025-08</date>
<extent unit="page">
<start>19</start>
<end>26</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T A Joint Multitask Model for Morpho-Syntactic Parsing
%A Améstica, Demian Inostroza
%A Mistica, Meladel
%A Vylomova, Ekaterina
%A Guest, Chris
%A Kurniawan, Kemal
%Y Goldman, Omer
%Y Weissweiler, Leonie
%Y Tsarfaty, Reut
%S Proceedings of The UniDive 2025 Shared Task on Multilingual Morpho-Syntactic Parsing
%D 2025
%8 August
%I Association for Computational Linguistics
%C Ljubljana, Slovenia
%@ 979-8-89176-320-3
%F amestica-etal-2025-joint
%X We present a joint multitask model for the Uni-Dive 2025 Morpho-Syntactic Parsing shared task, where systems predict both morphological and syntactic analyses following novel UD annotation scheme. Our system uses a shared XLM-RoBERTa encoder with three specialized decoders for content word identification, dependency parsing, and morphosyntactic feature prediction. Our model achieves the best overall performance on the shared task’s leaderboard covering nine typologically diverse languages, with an average MSLAS score of 78.7%, LAS of 80.1%, and Feats F1 of 90.3%. Our ablation studies show that matching the task’s gold tokenization and content word identification are crucial to model performance. Error analysis reveals that our model struggles with core grammatical cases (particularly Nom–Acc) and nominal features across languages.
%U https://aclanthology.org/2025.unidive-1.2/
%P 19-26
Markdown (Informal)
[A Joint Multitask Model for Morpho-Syntactic Parsing](https://aclanthology.org/2025.unidive-1.2/) (Améstica et al., UNIDIVE-SyntaxFest 2025)
ACL
- Demian Inostroza Améstica, Meladel Mistica, Ekaterina Vylomova, Chris Guest, and Kemal Kurniawan. 2025. A Joint Multitask Model for Morpho-Syntactic Parsing. In Proceedings of The UniDive 2025 Shared Task on Multilingual Morpho-Syntactic Parsing, pages 19–26, Ljubljana, Slovenia. Association for Computational Linguistics.