LatentRAG: 基于潜在空间的高效代理RAG推理与检索
LatentRAG: Latent Reasoning and Retrieval for Efficient Agentic RAG
AI总结：
LatentRAG通过将推理与检索从离散语言空间转移到连续潜在空间，提升代理RAG的效率，实验表明其在七种基准数据集上性能与显式方法相当，推理延迟降低约90%。
摘要：
单步检索增强生成（RAG）为简单问答任务提供了高效的外部信息整合方式，但难以处理复杂问题。代理RAG通过将单步检索替换为多步过程，使大语言模型（LLM）作为搜索代理生成中间思考和子查询以迭代交互。此过程因自回归生成长思考和子查询而产生显著延迟。为解决此限制，我们提出LatentRAG，一种将推理与检索从离散语言空间转移到连续潜在空间的新框架。与现有显式方法生成自然语言思考或子查询token-by-token不同，LatentRAG直接从隐藏状态生成潜在token。我们对LLM与密集检索模型在潜在空间对齐，实现对潜在子查询token的检索并支持端到端联合优化。为提高透明度并鼓励语义有意义的潜在表示，我们引入并行潜在解码机制将潜在token翻译回自然语言。在七种基准数据集上的大量实验表明，LatentRAG在性能上与显式代理RAG方法相当，同时将推理延迟降低约90%，显著缩小了与传统单步RAG的延迟差距。
英文摘要：
Single-step retrieval-augmented generation (RAG) provides an efficient way to incorporate external information for simple question answering tasks but struggles with complex questions. Agentic RAG extends this paradigm by replacing single-step retrieval with a multi-step process, in which the large language model (LLM) acts as a search agent that generates intermediate thoughts and subqueries to iteratively interact with the retrieval system. This iterative process incurs substantial latency due to the autoregressive generation of lengthy thoughts and subqueries. To address this limitation, we propose LatentRAG, a novel framework that shifts both reasoning and retrieval from discrete language space to continuous latent space. Unlike existing explicit methods that generate natural language thoughts or subqueries token-by-token, LatentRAG produces latent tokens for thoughts and subqueries directly from the hidden states in a single forward pass. We align LLMs with dense retrieval models in the latent space, enabling retrieval over latent subquery tokens and supporting end-to-end joint optimization. To improve transparency and encourage semantically meaningful latent representations, we incorporate a parallel latent decoding mechanism that translates latent tokens back into natural language. Extensive experiments on seven benchmark datasets show that LatentRAG achieves performance comparable to explicit agentic RAG methods while reducing inference latency by approximately 90%, substantially narrowing the latency gap with traditional single-step RAG.
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。
