DRAGIN: Dynamic Retrieval Augmented Generation based on the Information Needs of Large Language Models

TL;DR

DRAGIN introduces a dynamic retrieval augmented generation framework that adaptively determines when and what information to retrieve during LLM text generation, improving performance on knowledge-intensive tasks.

Contribution

The paper proposes DRAGIN, a novel framework for real-time, adaptive retrieval decisions in LLMs, addressing limitations of static retrieval strategies.

Findings

DRAGIN outperforms existing methods on four knowledge-intensive datasets.

Dynamic retrieval decisions improve LLM generation quality.

Open-sourced code and models facilitate further research.

Abstract

Dynamic retrieval augmented generation (RAG) paradigm actively decides when and what to retrieve during the text generation process of Large Language Models (LLMs). There are two key elements of this paradigm: identifying the optimal moment to activate the retrieval module (deciding when to retrieve) and crafting the appropriate query once retrieval is triggered (determining what to retrieve). However, current dynamic RAG methods fall short in both aspects. Firstly, the strategies for deciding when to retrieve often rely on static rules. Moreover, the strategies for deciding what to retrieve typically limit themselves to the LLM's most recent sentence or the last few tokens, while the LLM's real-time information needs may span across the entire context. To overcome these limitations, we introduce a new framework, DRAGIN, i.e., Dynamic Retrieval Augmented Generation based on the real-time Information Needs of LLMs. Our framework is specifically designed to make decisions on when and what to retrieve based on the LLM's real-time information needs during the text generation process. We evaluate DRAGIN along with existing methods comprehensively over 4 knowledge-intensive generation datasets. Experimental results show that DRAGIN achieves superior performance on all tasks, demonstrating the effectiveness of our method. We have open-sourced all the code, data, and models in GitHub: https://github.com/oneal2000/DRAGIN/tree/main