SEER: Self-Aligned Evidence Extraction for Retrieval-Augmented Generation
This addresses the problem of inefficient and low-quality evidence extraction in RAG systems for AI and NLP applications, representing an incremental improvement over existing heuristic-based methods.
The paper tackles the challenge of extracting evidence from retrieved passages in Retrieval-Augmented Generation (RAG) by proposing SEER, a self-aligned learning framework that optimizes an evidence extractor, resulting in improved RAG performance, enhanced faithfulness, helpfulness, and conciseness of evidence, and a 9.25 times reduction in evidence length.
Recent studies in Retrieval-Augmented Generation (RAG) have investigated extracting evidence from retrieved passages to reduce computational costs and enhance the final RAG performance, yet it remains challenging. Existing methods heavily rely on heuristic-based augmentation, encountering several issues: (1) Poor generalization due to hand-crafted context filtering; (2) Semantics deficiency due to rule-based context chunking; (3) Skewed length due to sentence-wise filter learning. To address these issues, we propose a model-based evidence extraction learning framework, SEER, optimizing a vanilla model as an evidence extractor with desired properties through self-aligned learning. Extensive experiments show that our method largely improves the final RAG performance, enhances the faithfulness, helpfulness, and conciseness of the extracted evidence, and reduces the evidence length by 9.25 times. The code will be available at https://github.com/HITsz-TMG/SEER.