SeaKR: Self-aware Knowledge Retrieval for Adaptive Retrieval Augmented Generation

TL;DR

SeaKR is an adaptive retrieval-augmented generation model that leverages LLMs' self-aware uncertainty to selectively retrieve, re-rank, and choose reasoning strategies, improving performance on question answering tasks.

Contribution

Introduces SeaKR, a novel adaptive RAG model that uses LLMs' self-aware uncertainty for retrieval, re-ranking, and reasoning strategy selection.

Findings

Outperforms existing adaptive RAG methods on QA datasets

Effectively utilizes self-aware uncertainty for retrieval and reasoning

Enhances complex task performance with multiple retrievals

Abstract

This paper introduces Self-aware Knowledge Retrieval (SeaKR), a novel adaptive RAG model that extracts self-aware uncertainty of LLMs from their internal states. SeaKR activates retrieval when the LLMs present high self-aware uncertainty for generation. To effectively integrate retrieved knowledge snippets, SeaKR re-ranks them based on LLM's self-aware uncertainty to preserve the snippet that reduces their uncertainty to the utmost. To facilitate solving complex tasks that require multiple retrievals, SeaKR utilizes their self-aware uncertainty to choose among different reasoning strategies. Our experiments on both complex and simple Question Answering datasets show that SeaKR outperforms existing adaptive RAG methods. We release our code at https://github.com/THU-KEG/SeaKR.