@inproceedings{ogezi-shi-2025-spare,
title = "{S}pa{RE}: Enhancing Spatial Reasoning in Vision-Language Models with Synthetic Data",
author = "Ogezi, Michael and
Shi, Freda",
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.387/",
doi = "10.18653/v1/2025.acl-long.387",
pages = "7855--7875",
ISBN = "979-8-89176-251-0",
abstract = "Vision-language models (VLMs) work well in tasks ranging from image captioning to visual question answering (VQA), yet they struggle with spatial reasoning, a key skill for understanding our physical world that humans excel at. We find that spatial relations are generally rare in widely used VL datasets, with only a few being well represented, while most form a long tail of underrepresented relations. This gap leaves VLMs ill-equipped to handle diverse spatial relationships. To bridge it, we construct a synthetic VQA dataset focused on spatial reasoning generated from hyper-detailed image descriptions in Localized Narratives, DOCCI, and PixMo-Cap. Our dataset consists of 455k samples containing 3.4 million QA pairs. Trained on this dataset, our Spatial-Reasoning Enhanced (SpaRE) VLMs show strong improvements on spatial reasoning benchmarks, achieving up to a 49{\%} performance gain on the What{'}s Up benchmark, while maintaining strong results on general tasks. Our work narrows the gap between human and VLM spatial reasoning and makes VLMs more capable in real-world tasks such as robotics and navigation. We plan to share our code and dataset in due course."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="ogezi-shi-2025-spare">
<titleInfo>
<title>SpaRE: Enhancing Spatial Reasoning in Vision-Language Models with Synthetic Data</title>
</titleInfo>
<name type="personal">
<namePart type="given">Michael</namePart>
<namePart type="family">Ogezi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Freda</namePart>
<namePart type="family">Shi</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>Vision-language models (VLMs) work well in tasks ranging from image captioning to visual question answering (VQA), yet they struggle with spatial reasoning, a key skill for understanding our physical world that humans excel at. We find that spatial relations are generally rare in widely used VL datasets, with only a few being well represented, while most form a long tail of underrepresented relations. This gap leaves VLMs ill-equipped to handle diverse spatial relationships. To bridge it, we construct a synthetic VQA dataset focused on spatial reasoning generated from hyper-detailed image descriptions in Localized Narratives, DOCCI, and PixMo-Cap. Our dataset consists of 455k samples containing 3.4 million QA pairs. Trained on this dataset, our Spatial-Reasoning Enhanced (SpaRE) VLMs show strong improvements on spatial reasoning benchmarks, achieving up to a 49% performance gain on the What’s Up benchmark, while maintaining strong results on general tasks. Our work narrows the gap between human and VLM spatial reasoning and makes VLMs more capable in real-world tasks such as robotics and navigation. We plan to share our code and dataset in due course.</abstract>
<identifier type="citekey">ogezi-shi-2025-spare</identifier>
<identifier type="doi">10.18653/v1/2025.acl-long.387</identifier>
<location>
<url>https://aclanthology.org/2025.acl-long.387/</url>
</location>
<part>
<date>2025-07</date>
<extent unit="page">
<start>7855</start>
<end>7875</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T SpaRE: Enhancing Spatial Reasoning in Vision-Language Models with Synthetic Data
%A Ogezi, Michael
%A Shi, Freda
%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 ogezi-shi-2025-spare
%X Vision-language models (VLMs) work well in tasks ranging from image captioning to visual question answering (VQA), yet they struggle with spatial reasoning, a key skill for understanding our physical world that humans excel at. We find that spatial relations are generally rare in widely used VL datasets, with only a few being well represented, while most form a long tail of underrepresented relations. This gap leaves VLMs ill-equipped to handle diverse spatial relationships. To bridge it, we construct a synthetic VQA dataset focused on spatial reasoning generated from hyper-detailed image descriptions in Localized Narratives, DOCCI, and PixMo-Cap. Our dataset consists of 455k samples containing 3.4 million QA pairs. Trained on this dataset, our Spatial-Reasoning Enhanced (SpaRE) VLMs show strong improvements on spatial reasoning benchmarks, achieving up to a 49% performance gain on the What’s Up benchmark, while maintaining strong results on general tasks. Our work narrows the gap between human and VLM spatial reasoning and makes VLMs more capable in real-world tasks such as robotics and navigation. We plan to share our code and dataset in due course.
%R 10.18653/v1/2025.acl-long.387
%U https://aclanthology.org/2025.acl-long.387/
%U https://doi.org/10.18653/v1/2025.acl-long.387
%P 7855-7875
Markdown (Informal)
[SpaRE: Enhancing Spatial Reasoning in Vision-Language Models with Synthetic Data](https://aclanthology.org/2025.acl-long.387/) (Ogezi & Shi, ACL 2025)
ACL