Relevance to Utility: Process-Supervised Rewrite for RAG
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866909983565676544 |
|---|---|
| author | Kim, Jaeyoung Kim, Jongho Hwang, Seung-won Song, Seoho Song, Young-In |
| author_facet | Kim, Jaeyoung Kim, Jongho Hwang, Seung-won Song, Seoho Song, Young-In |
| contents | Retrieval-augmented generation systems often suffer from a gap between optimizing retrieval relevance and generative utility. With such a gap, retrieved documents may be topically relevant but still lack the content needed for effective reasoning during generation. While existing bridge modules attempt to rewrite the retrieved text for better generation, we show how they fail by not capturing "document utility". In this work, we propose R2U, with a key distinction of approximating true utility through joint observation of rewriting and answering in the reasoning process. To distill, R2U scale such supervision to enhance reliability in distillation. We further construct utility-improvement supervision by measuring the generator's gain of the answer under the rewritten context, yielding signals for fine-tuning and preference optimization. We evaluate our method across multiple open-domain question-answering benchmarks. The empirical results demonstrate consistent improvements over strong bridging baselines |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_15577 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Relevance to Utility: Process-Supervised Rewrite for RAG Kim, Jaeyoung Kim, Jongho Hwang, Seung-won Song, Seoho Song, Young-In Computation and Language Artificial Intelligence Retrieval-augmented generation systems often suffer from a gap between optimizing retrieval relevance and generative utility. With such a gap, retrieved documents may be topically relevant but still lack the content needed for effective reasoning during generation. While existing bridge modules attempt to rewrite the retrieved text for better generation, we show how they fail by not capturing "document utility". In this work, we propose R2U, with a key distinction of approximating true utility through joint observation of rewriting and answering in the reasoning process. To distill, R2U scale such supervision to enhance reliability in distillation. We further construct utility-improvement supervision by measuring the generator's gain of the answer under the rewritten context, yielding signals for fine-tuning and preference optimization. We evaluate our method across multiple open-domain question-answering benchmarks. The empirical results demonstrate consistent improvements over strong bridging baselines |
| title | Relevance to Utility: Process-Supervised Rewrite for RAG |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2509.15577 |