@inproceedings{lin-etal-2025-v,
title = "{V}-{ALPHASOCIAL}: Benchmark and Self-Reflective Chain-of-Thought Generation for Visual Social Commonsense Reasoning",
author = "Lin, Zongyu and
Xu, Zhikun and
Song, Xiaohan and
Wan, Yixin and
Yao, Xingcheng and
Lin, Tsung-Han and
Song, Selina and
Subbaraman, Pranav and
Zhou, Ben and
Chang, Kai-Wei and
Sun, Yizhou",
editor = "Che, Wanxiang and
Nabende, Joyce and
Shutova, Ekaterina and
Pilehvar, Mohammad Taher",
booktitle = "Findings of the Association for Computational Linguistics: ACL 2025",
month = jul,
year = "2025",
address = "Vienna, Austria",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.findings-acl.975/",
doi = "10.18653/v1/2025.findings-acl.975",
pages = "19025--19047",
ISBN = "979-8-89176-256-5",
abstract = "Social commonsense reasoning naturally involves both the verbal and non-verbal cues of a social interaction. It is important for Large Vision-Language Models (VLMs) to leverage both textual and visual information in performing tasks like social understanding and reasoning. However, while current LLMs have shown good social reasoning capabilities in textual context, whether they can effectively incorporate visual information in social comprehension remains under-explored. To narrow the gap, we first construct and propose a benchmark: V-Social, featuring well-aligned text and visual content, tailored to assess visual social commonsense for multimodal foundation models. Through experimenting with V-Social, we find that even the most advanced VLM, GPT-4o, often falls short in social commonsense reasoning. This highlights the critical need to enhance the social grounding of VLMs. One major obstacle for improving this is the lack of high-quality data with good reasoning process. To overcome this obstacle, we introduce V-AlphaSocial, a novel method that generates high-quality chain-of-thought reasoning paths from unlabeled data. We design a visual reasoning reward model to improve VLM, and then iteratively refine both the VLM and the reward model. Our extensive analysis showcases how our method enhances social commonsense reasoning, proposing an effective approach that facilitates deeper exploration into field."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="lin-etal-2025-v">
<titleInfo>
<title>V-ALPHASOCIAL: Benchmark and Self-Reflective Chain-of-Thought Generation for Visual Social Commonsense Reasoning</title>
</titleInfo>
<name type="personal">
<namePart type="given">Zongyu</namePart>
<namePart type="family">Lin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhikun</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiaohan</namePart>
<namePart type="family">Song</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yixin</namePart>
<namePart type="family">Wan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xingcheng</namePart>
<namePart type="family">Yao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tsung-Han</namePart>
<namePart type="family">Lin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Selina</namePart>
<namePart type="family">Song</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Pranav</namePart>
<namePart type="family">Subbaraman</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ben</namePart>
<namePart type="family">Zhou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kai-Wei</namePart>
<namePart type="family">Chang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yizhou</namePart>
<namePart type="family">Sun</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>Findings of the Association for Computational Linguistics: ACL 2025</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">979-8-89176-256-5</identifier>
</relatedItem>
<abstract>Social commonsense reasoning naturally involves both the verbal and non-verbal cues of a social interaction. It is important for Large Vision-Language Models (VLMs) to leverage both textual and visual information in performing tasks like social understanding and reasoning. However, while current LLMs have shown good social reasoning capabilities in textual context, whether they can effectively incorporate visual information in social comprehension remains under-explored. To narrow the gap, we first construct and propose a benchmark: V-Social, featuring well-aligned text and visual content, tailored to assess visual social commonsense for multimodal foundation models. Through experimenting with V-Social, we find that even the most advanced VLM, GPT-4o, often falls short in social commonsense reasoning. This highlights the critical need to enhance the social grounding of VLMs. One major obstacle for improving this is the lack of high-quality data with good reasoning process. To overcome this obstacle, we introduce V-AlphaSocial, a novel method that generates high-quality chain-of-thought reasoning paths from unlabeled data. We design a visual reasoning reward model to improve VLM, and then iteratively refine both the VLM and the reward model. Our extensive analysis showcases how our method enhances social commonsense reasoning, proposing an effective approach that facilitates deeper exploration into field.</abstract>
<identifier type="citekey">lin-etal-2025-v</identifier>
<identifier type="doi">10.18653/v1/2025.findings-acl.975</identifier>
<location>
<url>https://aclanthology.org/2025.findings-acl.975/</url>
</location>
<part>
<date>2025-07</date>
<extent unit="page">
<start>19025</start>
<end>19047</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T V-ALPHASOCIAL: Benchmark and Self-Reflective Chain-of-Thought Generation for Visual Social Commonsense Reasoning
%A Lin, Zongyu
%A Xu, Zhikun
%A Song, Xiaohan
%A Wan, Yixin
%A Yao, Xingcheng
%A Lin, Tsung-Han
%A Song, Selina
%A Subbaraman, Pranav
%A Zhou, Ben
%A Chang, Kai-Wei
%A Sun, Yizhou
%Y Che, Wanxiang
%Y Nabende, Joyce
%Y Shutova, Ekaterina
%Y Pilehvar, Mohammad Taher
%S Findings of the Association for Computational Linguistics: ACL 2025
%D 2025
%8 July
%I Association for Computational Linguistics
%C Vienna, Austria
%@ 979-8-89176-256-5
%F lin-etal-2025-v
%X Social commonsense reasoning naturally involves both the verbal and non-verbal cues of a social interaction. It is important for Large Vision-Language Models (VLMs) to leverage both textual and visual information in performing tasks like social understanding and reasoning. However, while current LLMs have shown good social reasoning capabilities in textual context, whether they can effectively incorporate visual information in social comprehension remains under-explored. To narrow the gap, we first construct and propose a benchmark: V-Social, featuring well-aligned text and visual content, tailored to assess visual social commonsense for multimodal foundation models. Through experimenting with V-Social, we find that even the most advanced VLM, GPT-4o, often falls short in social commonsense reasoning. This highlights the critical need to enhance the social grounding of VLMs. One major obstacle for improving this is the lack of high-quality data with good reasoning process. To overcome this obstacle, we introduce V-AlphaSocial, a novel method that generates high-quality chain-of-thought reasoning paths from unlabeled data. We design a visual reasoning reward model to improve VLM, and then iteratively refine both the VLM and the reward model. Our extensive analysis showcases how our method enhances social commonsense reasoning, proposing an effective approach that facilitates deeper exploration into field.
%R 10.18653/v1/2025.findings-acl.975
%U https://aclanthology.org/2025.findings-acl.975/
%U https://doi.org/10.18653/v1/2025.findings-acl.975
%P 19025-19047
Markdown (Informal)
[V-ALPHASOCIAL: Benchmark and Self-Reflective Chain-of-Thought Generation for Visual Social Commonsense Reasoning](https://aclanthology.org/2025.findings-acl.975/) (Lin et al., Findings 2025)
ACL
- Zongyu Lin, Zhikun Xu, Xiaohan Song, Yixin Wan, Xingcheng Yao, Tsung-Han Lin, Selina Song, Pranav Subbaraman, Ben Zhou, Kai-Wei Chang, and Yizhou Sun. 2025. V-ALPHASOCIAL: Benchmark and Self-Reflective Chain-of-Thought Generation for Visual Social Commonsense Reasoning. In Findings of the Association for Computational Linguistics: ACL 2025, pages 19025–19047, Vienna, Austria. Association for Computational Linguistics.