H+嵌入：利用上下文相关短语协调全局与词元级检索
H+ Embedding: Harmonizing Global and Token-Level Retrieval with Context-Dependent Phrases
浏览论文内容
中文总结 AI 辅助
该研究提出H+嵌入多粒度检索器，以上下文相关短语为中间检索单元，在16项任务中提升检索效果，且降低了向量使用成本，为实用检索系统提供了质量与成本的平衡方案。
中文摘要 AI 辅助
术语密集型检索，尤其是医疗场景中的检索，依赖于保留多词实体、缩写、数值约束和组合概念。然而，现有表征处于两个极端：单向量检索器常过度压缩局部相关性信号，而词元级后期交互则保留了分词器的每个子词，带来了巨大的索引、存储和评分成本。这种不匹配引发了一个自然问题：上下文相关短语能否提供介于全局向量和词元之间的有用检索单元？我们提出H+嵌入，这是一种统一的多粒度检索器，可预测可变长度的短语划分，将未覆盖的词元保留为单例，并应用重要性引导的单元选择与加权MaxSim交互。在16项科学、医疗和双语任务中，其短语检索分支的宏观nDCG@10比全局检索分支高出6.91。它还几乎与Token级检索效果相当，同时使用的文档向量减少了13.7%，在中等向量预算下优于与内容无关的分组规则。因此，上下文相关短语交互为实用检索系统提供了介于全局压缩和词元级交互之间的中间质量-成本平衡点。
英文摘要
Terminology-intensive retrieval, especially in medical settings, depends on preserving multi-word entities, abbreviations, numerical constraints, and compositional concepts. However, existing representations lie at two extremes: single-vector retrievers often over-compress local relevance signals, while token-level late interaction retains every tokenizer subword at substantial indexing, storage, and scoring cost. This mismatch raises a natural question: can context-dependent phrases provide a useful retrieval unit between global vectors and tokens? We introduce H+ Embedding, a unified multi-granularity retriever that predicts variable-length phrase partitions, preserves uncovered tokens as singletons, and applies importance-guided unit selection with weighted MaxSim interaction. Across 16 scientific, medical, and bilingual tasks, its phrase retrieval branch exceeds the global retrieval branch by 6.91 macro nDCG@10. It also nearly matches Token while using 13.7% fewer document vectors and outperforms content-independent grouping rules under moderate vector budgets. Context-dependent phrase interaction therefore provides an intermediate quality-cost point between global compression and token-level interaction for practical retrieval systems.
