@inproceedings{kim-etal-2026-relevance,
title = "Relevance to Utility: Process-Supervised Rewrite for {RAG}",
author = "Kim, Jaeyoung and
Kim, Jongho and
Hwang, Seung-won and
Song, Seoho and
Song, Young-In",
editor = "Liakata, Maria and
Moreira, Viviane P. and
Zhang, Jiajun and
Jurgens, David",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
month = jul,
year = "2026",
address = "San Diego, California, United States",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-acl.1513/",
doi = "10.18653/v1/2026.findings-acl.1513",
pages = "30274--30293",
ISBN = "979-8-89176-395-1",
abstract = "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 this observation reliably, R2U scales 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."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="kim-etal-2026-relevance">
<titleInfo>
<title>Relevance to Utility: Process-Supervised Rewrite for RAG</title>
</titleInfo>
<name type="personal">
<namePart type="given">Jaeyoung</namePart>
<namePart type="family">Kim</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jongho</namePart>
<namePart type="family">Kim</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Seung-won</namePart>
<namePart type="family">Hwang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Seoho</namePart>
<namePart type="family">Song</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Young-In</namePart>
<namePart type="family">Song</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: ACL 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Maria</namePart>
<namePart type="family">Liakata</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Viviane</namePart>
<namePart type="given">P</namePart>
<namePart type="family">Moreira</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiajun</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">David</namePart>
<namePart type="family">Jurgens</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">San Diego, California, United States</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-395-1</identifier>
</relatedItem>
<abstract>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 this observation reliably, R2U scales 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.</abstract>
<identifier type="citekey">kim-etal-2026-relevance</identifier>
<identifier type="doi">10.18653/v1/2026.findings-acl.1513</identifier>
<location>
<url>https://aclanthology.org/2026.findings-acl.1513/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>30274</start>
<end>30293</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Relevance to Utility: Process-Supervised Rewrite for RAG
%A Kim, Jaeyoung
%A Kim, Jongho
%A Hwang, Seung-won
%A Song, Seoho
%A Song, Young-In
%Y Liakata, Maria
%Y Moreira, Viviane P.
%Y Zhang, Jiajun
%Y Jurgens, David
%S Findings of the Association for Computational Linguistics: ACL 2026
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, United States
%@ 979-8-89176-395-1
%F kim-etal-2026-relevance
%X 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 this observation reliably, R2U scales 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.
%R 10.18653/v1/2026.findings-acl.1513
%U https://aclanthology.org/2026.findings-acl.1513/
%U https://doi.org/10.18653/v1/2026.findings-acl.1513
%P 30274-30293
Markdown (Informal)
[Relevance to Utility: Process-Supervised Rewrite for RAG](https://aclanthology.org/2026.findings-acl.1513/) (Kim et al., Findings 2026)
ACL
- Jaeyoung Kim, Jongho Kim, Seung-won Hwang, Seoho Song, and Young-In Song. 2026. Relevance to Utility: Process-Supervised Rewrite for RAG. In Findings of the Association for Computational Linguistics: ACL 2026, pages 30274–30293, San Diego, California, United States. Association for Computational Linguistics.