arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2409.05152cs.CLcs.AIcs.DBcs.IRcs.LG
收藏

OneGen:面向 LLM 的高效单遍统一生成与检索

OneGen: Efficient One-Pass Unified Generation and Retrieval for LLMs

Zhejiang University(浙江大学);Ant Group(蚂蚁集团);Zhejiang University - Ant Group Joint Laboratory of Knowledge Graph(浙江大学-蚂蚁集团知识图谱联合实验室)

Jintian Zhang, Cheng Peng, Mengshu Sun, Xiang Chen, Lei Liang, Zhiqiang Zhang, Jun Zhou, Huajun Chen, Ningyu Zhang

AI总结:

本文提出 OneGen 框架,通过自回归生成检索 token,让单个 LLM 在一次前向传播中统一完成生成与向量检索,并在 RAG 和实体链接任务上验证其有效性与效率。

摘要:

尽管大语言模型(LLM)近期取得进展,显著增强了其在各类 NLP 任务上的生成能力,但 LLM 在直接处理检索任务方面仍存在局限。然而,许多实际应用要求将检索与生成无缝集成。本文提出一种新颖且高效的单遍生成与检索框架(OneGen),旨在提升 LLM 在需要同时进行生成和检索的任务上的表现。所提框架通过纳入以自回归方式生成的检索 token,弥合了传统上相互分离的生成与检索训练方法。这使单个 LLM 能够在统一的前向传播中同时处理两项任务。我们在两类不同的复合任务——RAG 和实体链接(Entity Linking)——上开展实验,以验证 OneGen 在训练和推理中的可插拔性、有效性与效率。此外,我们的结果表明,在同一上下文中集成生成与检索,能够在提升检索性能的同时保持 LLM 的生成能力。据我们所知,OneGen 是首个使 LLM 能够在生成过程中执行向量检索的方法。

英文摘要:

Despite the recent advancements in Large Language Models (LLMs), which have significantly enhanced the generative capabilities for various NLP tasks, LLMs still face limitations in directly handling retrieval tasks. However, many practical applications demand the seamless integration of both retrieval and generation. This paper introduces a novel and efficient One-pass Generation and retrieval framework (OneGen), designed to improve LLMs' performance on tasks that require both generation and retrieval. The proposed framework bridges the traditionally separate training approaches for generation and retrieval by incorporating retrieval tokens generated autoregressively. This enables a single LLM to handle both tasks simultaneously in a unified forward pass. We conduct experiments on two distinct types of composite tasks, RAG and Entity Linking, to validate the pluggability, effectiveness, and efficiency of OneGen in training and inference. Furthermore, our results show that integrating generation and retrieval within the same context preserves the generative capabilities of LLMs while improving retrieval performance. To the best of our knowledge, OneGen is the first to enable LLMs to conduct vector retrieval during the generation.

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。

↑