SeaKR: Self-aware Knowledge Retrieval for Adaptive Retrieval Augmented Generation
This addresses the problem of inefficient retrieval in RAG systems for AI researchers and practitioners, but it is incremental as it builds on existing adaptive RAG methods.
The paper tackles the problem of adaptive retrieval in RAG systems by introducing SeaKR, which uses LLMs' internal states to measure self-aware uncertainty for triggering retrieval and re-ranking knowledge snippets. The result shows SeaKR outperforms existing adaptive RAG methods on complex and simple Question Answering datasets.
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.