Hierarchical Document Refinement for Long-context Retrieval-augmented Generation
This work addresses efficiency and performance issues in long-text RAG applications, which is an incremental improvement for real-world scenarios.
The paper tackles the problem of redundant information and noise in long-context retrieval-augmented generation (RAG) applications, which leads to high inference costs and reduced performance, by proposing LongRefiner, an efficient plug-and-play refiner that achieves competitive performance on seven QA datasets while using 10x fewer computational costs and latency compared to the best baseline.
Real-world RAG applications often encounter long-context input scenarios, where redundant information and noise results in higher inference costs and reduced performance. To address these challenges, we propose LongRefiner, an efficient plug-and-play refiner that leverages the inherent structural characteristics of long documents. LongRefiner employs dual-level query analysis, hierarchical document structuring, and adaptive refinement through multi-task learning on a single foundation model. Experiments on seven QA datasets demonstrate that LongRefiner achieves competitive performance in various scenarios while using 10x fewer computational costs and latency compared to the best baseline. Further analysis validates that LongRefiner is scalable, efficient, and effective, providing practical insights for real-world long-text RAG applications. Our code is available at https://github.com/ignorejjj/LongRefiner.