用于密集检索的检索 grounding 潜在推理

中文摘要 AI 辅助

需要大量推理的检索任务要求文本表示不仅能捕捉语义相似性，还能捕捉在给定检索指令下判断相关性所需的推理能力。现有的推理增强嵌入模型通过将推理信息融入密集表示来提升检索效果，但其监督通常受最终检索目标主导，导致潜在推理轨迹可能学习到捷径推理模式，这些模式虽能保持检索性能，却无法产生有意义的增量检索增益。本文提出 Retrieval Grounding Latent Reasoning（RGLT），一种用于密集检索的潜在推理框架，它明确将中间潜在转换与检索改进关联起来。RGLT 通过由 silent tokens 构建的指令条件潜在推理轨迹在隐空间中执行非自回归推理，它结合过程监督的显式到隐式蒸馏与检索 grounding 监督，使用阶段式 CoT 重构塑造中间潜在状态，并利用检索效应信用优化潜在推理轨迹上的增量检索增益。在推理密集型检索基准上的实验表明，RGLT 在保持高效嵌入推理的同时，始终优于强大的基线模型。

英文摘要

Reasoning-intensive retrieval requires text representations to capture not only semantic similarity, but also the reasoning needed to determine relevance under a given retrieval instruction. Existing reasoning-enhanced embedding models improve retrieval by incorporating reasoning information into dense representations, yet their supervision is typically dominated by the final retrieval objective. As a result, latent reasoning trajectories may learn shortcut reasoning patterns that preserve retrieval performance without producing meaningful incremental retrieval gains. We propose Retrieval Grounding Latent Reasoning (RGLT), a latent reasoning framework for dense retrieval that explicitly connects intermediate latent transitions with retrieval improvements. RGLT performs non-autoregressive reasoning in hidden space through an instruction-conditioned latent reasoning trajectory constructed from silent tokens. It combines process-supervised explicit-to-implicit distillation with retrieval-grounded supervision, using stage-wise CoT reconstruction to shape intermediate latent states and retrieval-effect credit to optimize incremental retrieval gains across the latent reasoning trajectories. Experiments on reasoning-intensive retrieval benchmarks show that RGLT consistently outperforms strong baselines while preserving efficient embedding inference.

发表机构

• Meituan（美团）

机构由 AI 辅助整理，请以论文原文为准。