arXiv 2601.079862026-02-24cs.CLcs.CV

VULCA-Bench：一个多文化视觉-语言基准，用于评估文化理解

VULCA-Bench: A Multicultural Vision-Language Benchmark for Evaluating Cultural Understanding

DJCAD, University of Dundee, United Kingdom（邓迪大学DJCAD部门）；University of California, Santa Cruz, USA（加州大学圣克ruz分校）；Analogy AI；East China Normal University, China（华东师范大学）；Nanjing University, China（南京大学）；Department of Mechanical Engineering, School of Engineering, University of Birmingham（伯明翰大学工程学院机械工程系）

Haorui Yu, Diji Yang, Hang He, Fengrui Zhang, Qiufeng Yi

备注：: 8 pages, 4 figures, submitted to ACL 2026 Dataset Track

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

VULCA-Bench通过多文化艺术批评基准评估视觉-语言模型在文化理解上的能力，包含7,410对图像-批评对，涵盖八个文化传统，通过五层框架评估文化理解，揭示更高阶推理更具挑战性。

摘要：

我们介绍了VULCA-Bench，一个多文化艺术批评基准，用于评估视觉-语言模型（VLMs）在超越表层视觉感知之外的文化理解能力。现有VLM基准主要测量L1-L2能力（物体识别、场景描述和事实性问题回答），而低估了更高阶的文化解释能力。VULCA-Bench包含7,410对匹配的图像-批评对，涵盖八个文化传统，具有中英双语覆盖。我们通过五层框架（L1-L5，从视觉感知到哲学美学）来定义文化理解，具体化为225个文化特定维度，并通过专家撰写的双语批评来支持。我们的初步结果表明，更高层推理（L3-L5）比视觉和技术分析（L1-L2）更具挑战性。该数据集、评估脚本和注释工具可在CC BY 4.0下获取，网址为https://github.com/yha9806/VULCA-Bench。

英文摘要：

We introduce VULCA-Bench, a multicultural art-critique benchmark for evaluating Vision-Language Models' (VLMs) cultural understanding beyond surface-level visual perception. Existing VLM benchmarks predominantly measure L1-L2 capabilities (object recognition, scene description, and factual question answering) while under-evaluate higher-order cultural interpretation. VULCA-Bench contains 7,410 matched image-critique pairs spanning eight cultural traditions, with Chinese-English bilingual coverage. We operationalise cultural understanding using a five-layer framework (L1-L5, from Visual Perception to Philosophical Aesthetics), instantiated as 225 culture-specific dimensions and supported by expert-written bilingual critiques. Our pilot results indicate that higher-layer reasoning (L3-L5) is consistently more challenging than visual and technical analysis (L1-L2). The dataset, evaluation scripts, and annotation tools are available under CC BY 4.0 at https://github.com/yha9806/VULCA-Bench.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。