EviNote-RAG: Enhancing RAG Models via Answer-Supportive Evidence Notes
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911214880161792 |
|---|---|
| author | Dai, Yuqin Wang, Guoqing Wang, Yuan Dou, Kairan Zhou, Kaichen Zhang, Zhanwei Yang, Shuo Tang, Fei Yin, Jun Zeng, Pengyu Ying, Zhenzhe Yi, Can Meng, Changhua Zhou, Yuchen Shen, Yongliang Lu, Shuai |
| author_facet | Dai, Yuqin Wang, Guoqing Wang, Yuan Dou, Kairan Zhou, Kaichen Zhang, Zhanwei Yang, Shuo Tang, Fei Yin, Jun Zeng, Pengyu Ying, Zhenzhe Yi, Can Meng, Changhua Zhou, Yuchen Shen, Yongliang Lu, Shuai |
| contents | Retrieval-Augmented Generation (RAG) has advanced open-domain question answering by incorporating external information into model reasoning. However, effectively leveraging external information to enhance reasoning presents the following challenges: (1) low signal-to-noise ratio, where answer-supportive external information is diluted by irrelevant material, and (2) error accumulation, which arises in multi-hop reasoning when incomplete or misleading information is incorporated. To address these challenges, we introduce EviNote-RAG, a framework that follows a retrieve-note-answer workflow. Instead of reasoning directly over raw external information, the model first produces Supportive-Evidence Notes (SENs), which concisely preserve answer-critical information and explicitly mark key and uncertainty information to improve accuracy. We further design an entailment-based Evidence Quality Reward (EQR) to ensure that SENs are logically sufficient to derive the final answer, thereby enhancing SENs' quality. Experiments on both in-domain and out-of-domain QA benchmarks show that EviNote-RAG achieves state-of-the-art performance, improving answer accuracy, training stability, robustness, and efficiency. In particular, it yields relative F1 gains of 20% on HotpotQA (+0.093), 40% on Bamboogle (+0.151), and 91% on 2Wiki (+0.256), benefiting from improvements in the reasoning process. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_00877 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | EviNote-RAG: Enhancing RAG Models via Answer-Supportive Evidence Notes Dai, Yuqin Wang, Guoqing Wang, Yuan Dou, Kairan Zhou, Kaichen Zhang, Zhanwei Yang, Shuo Tang, Fei Yin, Jun Zeng, Pengyu Ying, Zhenzhe Yi, Can Meng, Changhua Zhou, Yuchen Shen, Yongliang Lu, Shuai Computation and Language Retrieval-Augmented Generation (RAG) has advanced open-domain question answering by incorporating external information into model reasoning. However, effectively leveraging external information to enhance reasoning presents the following challenges: (1) low signal-to-noise ratio, where answer-supportive external information is diluted by irrelevant material, and (2) error accumulation, which arises in multi-hop reasoning when incomplete or misleading information is incorporated. To address these challenges, we introduce EviNote-RAG, a framework that follows a retrieve-note-answer workflow. Instead of reasoning directly over raw external information, the model first produces Supportive-Evidence Notes (SENs), which concisely preserve answer-critical information and explicitly mark key and uncertainty information to improve accuracy. We further design an entailment-based Evidence Quality Reward (EQR) to ensure that SENs are logically sufficient to derive the final answer, thereby enhancing SENs' quality. Experiments on both in-domain and out-of-domain QA benchmarks show that EviNote-RAG achieves state-of-the-art performance, improving answer accuracy, training stability, robustness, and efficiency. In particular, it yields relative F1 gains of 20% on HotpotQA (+0.093), 40% on Bamboogle (+0.151), and 91% on 2Wiki (+0.256), benefiting from improvements in the reasoning process. |
| title | EviNote-RAG: Enhancing RAG Models via Answer-Supportive Evidence Notes |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2509.00877 |