KARMA-MV：音乐视频上的因果问答基准
KARMA-MV: A Benchmark for Causal Question Answering on Music Videos
AI总结：
KARMA-MV是一个基于2682个YouTube音乐视频构建的大规模多选问答数据集，旨在测试模型整合时序音频视觉线索和视觉到音乐影响的能力，通过因果知识图谱方法提升音乐视频因果推理性能。
摘要：
尽管在视频问答和跨模态理解方面取得了显著进展，但关于视觉动态如何驱动音乐结构的因果推理在音乐视频中仍被低估。我们介绍了KARMA-MV，一个从2682个YouTube音乐视频衍生出的大型多选问答数据集，旨在测试模型整合时序音频视觉线索并推理视觉到音乐影响的能力。与需要人工标注的传统数据集不同，KARMA-MV利用LLM推理进行可扩展的生成和验证，产生37737个多选问题。我们提出了一种因果知识图谱（CKG）方法，通过结构化检索跨模态依赖性增强视觉语言模型（VLMs）。在最先进的VLMs和LLMs上的实验表明，CKG基础带来了持续的提升，尤其是对较小的模型，确立了显式因果结构在音乐视频推理中的价值。KARMA-MV为超越相关性的因果音频视觉理解提供了新的基准。
英文摘要：
While significant progress has been made in Video Question Answering and cross-modal understanding, causal reasoning about how visual dynamics drive musical structure in music videos remains under-explored. We introduce KARMA-MV, a large-scale multiple-choice QA dataset derived from 2,682 YouTube music videos, designed to test models' ability to integrate temporal audio-visual cues and reason about visual-to-musical influence across reasoning, prediction, and counterfactual questions. Unlike traditional datasets requiring manual annotation, KARMA-MV leverages LLM reasoning for scalable generation and validation, yielding 37,737 MCQs. We propose a causal knowledge graph (CKG) approach that augments vision-language models (VLMs) with structured retrieval of cross-modal dependencies. Experiments on state-of-the-art VLMs and LLMs show consistent gains from CKG grounding -- especially for smaller models -- establishing the value of explicit causal structure for music-video reasoning. KARMA-MV provides a new benchmark for advancing causal audio-visual understanding beyond correlation.
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。
