MUST-RAG: MUSical Text Question Answering with Retrieval Augmented Generation
MUST-RAG: 基于检索增强生成的音乐文本问答
机构 * Graduate School of Culture Technology, KAIST, South Korea(文化科技研究生院,韩国釜山科学技术院)
专题命中 检索器与排序 :retrieval augmented generation(title,abstract);RAG(title,abstract);分类 cs.IR、cs.CL、cs.AI
AI总结 MUST-RAG通过引入音乐专用数据库和检索增强生成技术,提升通用语言模型在音乐问答任务中的表现。
Comments This is an earlier version of the paper - ArtistMus: A Globally Diverse, Artist-Centric Benchmark for Retrieval-Augmented Music Question Answering. The latest version is available at: (arXiv:2512.05430)