发表机构
Seoul National University（首尔大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
该研究针对现有IR基准的不足，构建了多视角多领域多模态IR基准Multi³IR，提出SPIN方法提升检索器的视角覆盖度，实验验证其性能优势与泛化性。
AI 中文摘要
信息检索（IR）越来越多地针对允许多样化视角的开放式查询。然而，现有的IR基准主要聚焦于封闭式查询，即便开放式基准也大多包含其支持文档仅覆盖单一主题领域和模态的查询。我们推出Multi³IR，这是一个评估检索器在跨不同领域和模态的开放式查询中覆盖多方面视角能力的基准。它包含104.9K条Stack Exchange查询，每条都带有捕捉查询隐含观点的视角描述。我们进一步提出SPIN，这是一种参数和标签高效的方法，用于学习噪声向量以将嵌入引导至多样且有意义的语义方向。实验表明，现有的多模态检索器存在单视角偏差，而SPIN在Multi³IR上大幅提升了视角覆盖度，并且能很好地泛化到未见过的开放式IR基准。数据集和实验代码可在该https URL获取。
英文摘要
Information retrieval (IR) increasingly targets open-ended queries that admit diverse perspectives. Existing IR benchmarks, however, focus primarily on closed-ended queries, while even open-ended benchmarks largely consist of queries whose supporting documents span a single subject domain and modality. We introduce Multi$^3$IR, a benchmark that evaluates how well retrievers cover the multifaceted perspectives of open-ended queries across diverse domains and modalities. It comprises 104.9K Stack Exchange queries, each annotated with perspective descriptions that capture the query's implicit viewpoints. We further propose SPIN, a parameter- and label-efficient method that learns noise vectors to steer embeddings toward diverse yet meaningful semantic directions. Experiments show that existing multimodal retrievers suffer from single-perspective bias, while SPIN substantially improves perspective coverage on Multi$^3$IR and generalizes well to unseen open-ended IR benchmarks. The dataset and experimental code are available at https://github.com/seokwon99/Multi3IR.
CommentsEMNLP 2026; code is available at https://github.com/seokwon99/Multi3IR
