SEER: Self-Aligned Evidence Extraction for Retrieval-Augmented Generation

TL;DR

SEER introduces a self-aligned learning framework for evidence extraction in retrieval-augmented generation, significantly improving evidence quality and reducing length compared to heuristic methods.

Contribution

The paper presents a novel model-based evidence extraction method that overcomes limitations of heuristic approaches through self-aligned learning, enhancing RAG performance.

Findings

Improves RAG performance significantly

Enhances faithfulness, helpfulness, and conciseness of evidence

Reduces evidence length by 9.25 times

Abstract

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.