@inproceedings{zhang-etal-2025-exploring,
title = "Exploring Multimodal Relation Extraction of Hierarchical Tabular Data with Multi-task Learning",
author = "Zhang, Xinyu and
Song, Aibo and
Qiu, Jingyi and
Jin, Jiahui and
Zhang, Tianbo and
Fang, Xiaolin",
editor = "Che, Wanxiang and
Nabende, Joyce and
Shutova, Ekaterina and
Pilehvar, Mohammad Taher",
booktitle = "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2025",
address = "Vienna, Austria",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.acl-long.1298/",
doi = "10.18653/v1/2025.acl-long.1298",
pages = "26770--26781",
ISBN = "979-8-89176-251-0",
abstract = "Relation Extraction (RE) is a key task in table understanding, aiming to extract semantic relations between columns. However, complex tables with hierarchical headers are hard to obtain high-quality textual formats (e.g., Markdown) for input under practical scenarios like webpage screenshots and scanned documents, while table images are more accessible and intuitive. Besides, existing works overlook the need of mining relations among multiple columns rather than just the semantic relation between two specific columns in real-world practice. In this work, we explore utilizing Multimodal Large Language Models (MLLMs) to address RE in tables with complex structures. We creatively extend the concept of RE to include calculational relations, enabling multi-task learning of both semantic and calculational RE for mutual reinforcement. Specifically, we reconstruct table images into graph structure based on neighboring nodes to extract graph-level visual features. Such feature enhancement alleviates the insensitivity of MLLMs to the positional information within table images. We then propose a Chain-of-Thought distillation framework with self-correction mechanism to enhance MLLMs' reasoning capabilities without increasing parameter scale. Our method significantly outperforms most baselines on wide datasets. Additionally, we release a benchmark dataset for calculational RE in complex tables."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zhang-etal-2025-exploring">
<titleInfo>
<title>Exploring Multimodal Relation Extraction of Hierarchical Tabular Data with Multi-task Learning</title>
</titleInfo>
<name type="personal">
<namePart type="given">Xinyu</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Aibo</namePart>
<namePart type="family">Song</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jingyi</namePart>
<namePart type="family">Qiu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiahui</namePart>
<namePart type="family">Jin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tianbo</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiaolin</namePart>
<namePart type="family">Fang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Wanxiang</namePart>
<namePart type="family">Che</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Joyce</namePart>
<namePart type="family">Nabende</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ekaterina</namePart>
<namePart type="family">Shutova</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mohammad</namePart>
<namePart type="given">Taher</namePart>
<namePart type="family">Pilehvar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Vienna, Austria</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-251-0</identifier>
</relatedItem>
<abstract>Relation Extraction (RE) is a key task in table understanding, aiming to extract semantic relations between columns. However, complex tables with hierarchical headers are hard to obtain high-quality textual formats (e.g., Markdown) for input under practical scenarios like webpage screenshots and scanned documents, while table images are more accessible and intuitive. Besides, existing works overlook the need of mining relations among multiple columns rather than just the semantic relation between two specific columns in real-world practice. In this work, we explore utilizing Multimodal Large Language Models (MLLMs) to address RE in tables with complex structures. We creatively extend the concept of RE to include calculational relations, enabling multi-task learning of both semantic and calculational RE for mutual reinforcement. Specifically, we reconstruct table images into graph structure based on neighboring nodes to extract graph-level visual features. Such feature enhancement alleviates the insensitivity of MLLMs to the positional information within table images. We then propose a Chain-of-Thought distillation framework with self-correction mechanism to enhance MLLMs’ reasoning capabilities without increasing parameter scale. Our method significantly outperforms most baselines on wide datasets. Additionally, we release a benchmark dataset for calculational RE in complex tables.</abstract>
<identifier type="citekey">zhang-etal-2025-exploring</identifier>
<identifier type="doi">10.18653/v1/2025.acl-long.1298</identifier>
<location>
<url>https://aclanthology.org/2025.acl-long.1298/</url>
</location>
<part>
<date>2025-07</date>
<extent unit="page">
<start>26770</start>
<end>26781</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Exploring Multimodal Relation Extraction of Hierarchical Tabular Data with Multi-task Learning
%A Zhang, Xinyu
%A Song, Aibo
%A Qiu, Jingyi
%A Jin, Jiahui
%A Zhang, Tianbo
%A Fang, Xiaolin
%Y Che, Wanxiang
%Y Nabende, Joyce
%Y Shutova, Ekaterina
%Y Pilehvar, Mohammad Taher
%S Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2025
%8 July
%I Association for Computational Linguistics
%C Vienna, Austria
%@ 979-8-89176-251-0
%F zhang-etal-2025-exploring
%X Relation Extraction (RE) is a key task in table understanding, aiming to extract semantic relations between columns. However, complex tables with hierarchical headers are hard to obtain high-quality textual formats (e.g., Markdown) for input under practical scenarios like webpage screenshots and scanned documents, while table images are more accessible and intuitive. Besides, existing works overlook the need of mining relations among multiple columns rather than just the semantic relation between two specific columns in real-world practice. In this work, we explore utilizing Multimodal Large Language Models (MLLMs) to address RE in tables with complex structures. We creatively extend the concept of RE to include calculational relations, enabling multi-task learning of both semantic and calculational RE for mutual reinforcement. Specifically, we reconstruct table images into graph structure based on neighboring nodes to extract graph-level visual features. Such feature enhancement alleviates the insensitivity of MLLMs to the positional information within table images. We then propose a Chain-of-Thought distillation framework with self-correction mechanism to enhance MLLMs’ reasoning capabilities without increasing parameter scale. Our method significantly outperforms most baselines on wide datasets. Additionally, we release a benchmark dataset for calculational RE in complex tables.
%R 10.18653/v1/2025.acl-long.1298
%U https://aclanthology.org/2025.acl-long.1298/
%U https://doi.org/10.18653/v1/2025.acl-long.1298
%P 26770-26781
Markdown (Informal)
[Exploring Multimodal Relation Extraction of Hierarchical Tabular Data with Multi-task Learning](https://aclanthology.org/2025.acl-long.1298/) (Zhang et al., ACL 2025)
ACL