将检索分歧作为信号：诊断与调控多语言稠密嵌入器
Retrieval Disagreement as a Signal: Diagnosing and Steering Multilingual Dense Embedders
AI总结：
该研究针对多语言稠密嵌入器在共享索引下的检索分歧问题，提出分歧引导调控方法，可减少同语言拥挤，在召回率上保持竞争力，与均值移除方法形成权衡。
摘要：
多语言稠密检索器将所有语言置于同一共享空间，使得问题无论以何种语言书写，都能找到相关文档。我们发现，用不同语言提出同一问题，会检索到差异显著的文档，且这种分歧并非源于语言理解能力薄弱：每种语言在自身场景下表现良好，且对候选文档的排序与其他语言一致；仅当语言共享索引时才会出现分歧。我们追溯其机制：每种语言维持自身的排序，但在共享索引中，其正确答案被压至靠后位置，被模型对应的查询语言或英语的不相关文档所掩盖。我们提出分歧引导调控方法：针对每种语言，对比其他语言对某一问题达成一致的文档与该语言特有的检索文档，将它们的表示差异作为特定语言方向，在推理时添加到该语言的查询中。与语言均值移除、子空间移除、单语言旋转等基线方法相比，该方法能最有效地减少同语言拥挤问题，同时在召回率上保持竞争力，与更好保留排序质量的均值移除方法形成权衡。
英文摘要：
Multilingual dense retrievers place all languages in one shared space so that a question finds relevant documents, regardless of which language they are written in. We show that the same question, asked in different languages, retrieves substantially different documents, and this disagreement is not because of weak language understanding. Each language performs well on its own and ranks candidates consistently with the other languages; divergence appears only when languages share an index. We trace the mechanism: each language keeps its own ranking, but in the shared index, its correct answers are pushed far down, buried under non-relevant documents that are either in the query's own language or in English, depending on the model. We introduce disagreement-guided steering: for each language, we contrast the documents that other languages agree on for a question against those that language retrieves idiosyncratically, and take the difference of their representations as a language-specific direction, added to that language's queries at inference. Against baselines of language-mean removal, subspace removal, and per-language rotation, it most strongly reduces same-language crowding while remaining competitive on recall, trading off against mean removal, which better preserves ranking quality.
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。
