LITTA：晚期交互与测试时对齐用于视觉接地多模态检索
LITTA: Late-Interaction and Test-Time Alignment for Visually-Grounded Multimodal Retrieval
AI总结：
LITTA通过查询扩展提升多模态文档检索，利用晚期交互评分和反向排名融合提高检索鲁棒性，适用于计算机科学、制药和工业手册等多领域。
摘要：
从视觉丰富文档（如教科书、技术报告和手册）中检索相关证据具有挑战性，因为上下文长、布局复杂且用户问题与支持页面之间词汇重叠弱。我们提出LITTA，一种以查询扩展为中心的证据页面检索框架，无需重新训练检索器即可提升多模态文档检索。给定用户查询，LITTA使用大型语言模型生成互补的查询变体，并使用冻结的视觉检索器和晚期交互评分检索每个变体的候选页面。通过反向排名融合聚合扩展查询的候选页面以提高证据覆盖率并减少对任何单一表达式的敏感性。这种简单的测试时策略显著提高了检索鲁棒性，同时与现有多模态嵌入索引兼容。我们在三个领域（计算机科学、制药和工业手册）的视觉接地文档检索任务上评估了LITTA。多查询检索在top-k准确率、召回率和MRR方面优于单查询检索，特别是在视觉和语义变化大的领域有显著提升。此外，精度-效率权衡可通过查询变体数量直接控制，使LITTA在延迟约束下具有实用性。这些结果表明，查询扩展提供了一种简单而有效的机制来提高视觉接地多模态检索。
英文摘要：
Retrieving relevant evidence from visually rich documents such as textbooks, technical reports, and manuals is challenging due to long context, complex layouts, and weak lexical overlap between user questions and supporting pages. We propose LITTA, a query-expansion-centric retrieval framework for evidence page retrieval that improves multimodal document retrieval without retriever retraining. Given a user query, LITTA generates complementary query variants using a large language model and retrieves candidate pages for each variant using a frozen vision retriever with late-interaction scoring. Candidates from expanded queries are then aggregated through reciprocal rank fusion to improve evidence coverage and reduce sensitivity to any single phrasing. This simple test-time strategy significantly improves retrieval robustness while remaining compatible with existing multimodal embedding indices. We evaluate LITTA on visually grounded document retrieval tasks across three domains: computer science, pharmaceuticals, and industrial manuals. Multi-query retrieval consistently improves top-k accuracy, recall, and MRR compared to single-query retrieval, with particularly large gains in domains with high visual and semantic variability. Moreover, the accuracy-efficiency trade-off is directly controllable by the number of query variants, making LITTA practical for deployment under latency constraints. These results demonstrate that query expansion provides a simple yet effective mechanism for improving visually grounded multimodal retrieval.
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。
