@inproceedings{yuan-etal-2026-r3a,
title = "{R}{\textthreesuperior}{A}: Reinforced Reasoning for Relevance Assessment for {RAG} in User-Generated Content Platforms",
author = "Yuan, Xiaowei and
Jin, Lei and
Zhang, Haoxin and
Huang, Ziyang and
Gao, Yan and
Yiwu and
Hu, Yao and
Zhao, Jun and
Liu, Kang",
editor = "Li, Yunyao and
Rehm, Georg and
Tu, Mei",
booktitle = "Proceedings of the 64th Annual Meeting of the {A}ssociation for {C}omputational {L}inguistics ({ACL} 2026)",
month = jul,
year = "2026",
address = "San Diego, California, USA",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.acl-industry.10/",
pages = "127--140",
ISBN = "979-8-89176-394-4",
abstract = "Retrieval-augmented generation (RAG) plays a critical role in user-generated content (UGC) platforms, but its effectiveness critically depends on accurate query{--}document relevance assessment. Despite recent advances in applying large language models (LLMs) to relevance modeling, UGC platforms present unique challenges: 1) ambiguous user intent due to sparse user feedback in RAG scenarios, and 2) asymmetric relevance, where relevance is driven by localized answer-bearing content rather than global query{--}document similarity. To address these issues, we propose the Reinforced Reasoning model for Relevance Assessment (R{\textthreesuperior}A), which decomposes relevance assessment into intent inference and evidence grounding. R{\textthreesuperior}A leverages auxiliary high-clicked documents to infer latent query intent, and extracts verbatim evidence fragments to ground relevance decisions, reducing noise sensitivity and improving asymmetric relevance modeling. Experimental results demonstrate that R{\textthreesuperior}A substantially outperforms strong baselines on offline benchmarks, while the distilled R{\textthreesuperior}A-1.5B model achieves significant gains in large-scale online A/B testing, effectively balancing performance and practical deployability."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="yuan-etal-2026-r3a">
<titleInfo>
<title>R³A: Reinforced Reasoning for Relevance Assessment for RAG in User-Generated Content Platforms</title>
</titleInfo>
<name type="personal">
<namePart type="given">Xiaowei</namePart>
<namePart type="family">Yuan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lei</namePart>
<namePart type="family">Jin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Haoxin</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ziyang</namePart>
<namePart type="family">Huang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yan</namePart>
<namePart type="family">Gao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name>
<namePart>Yiwu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yao</namePart>
<namePart type="family">Hu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jun</namePart>
<namePart type="family">Zhao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kang</namePart>
<namePart type="family">Liu</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>Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yunyao</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Georg</namePart>
<namePart type="family">Rehm</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mei</namePart>
<namePart type="family">Tu</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, USA</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-394-4</identifier>
</relatedItem>
<abstract>Retrieval-augmented generation (RAG) plays a critical role in user-generated content (UGC) platforms, but its effectiveness critically depends on accurate query–document relevance assessment. Despite recent advances in applying large language models (LLMs) to relevance modeling, UGC platforms present unique challenges: 1) ambiguous user intent due to sparse user feedback in RAG scenarios, and 2) asymmetric relevance, where relevance is driven by localized answer-bearing content rather than global query–document similarity. To address these issues, we propose the Reinforced Reasoning model for Relevance Assessment (R³A), which decomposes relevance assessment into intent inference and evidence grounding. R³A leverages auxiliary high-clicked documents to infer latent query intent, and extracts verbatim evidence fragments to ground relevance decisions, reducing noise sensitivity and improving asymmetric relevance modeling. Experimental results demonstrate that R³A substantially outperforms strong baselines on offline benchmarks, while the distilled R³A-1.5B model achieves significant gains in large-scale online A/B testing, effectively balancing performance and practical deployability.</abstract>
<identifier type="citekey">yuan-etal-2026-r3a</identifier>
<location>
<url>https://aclanthology.org/2026.acl-industry.10/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>127</start>
<end>140</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T R³A: Reinforced Reasoning for Relevance Assessment for RAG in User-Generated Content Platforms
%A Yuan, Xiaowei
%A Jin, Lei
%A Zhang, Haoxin
%A Huang, Ziyang
%A Gao, Yan
%A Hu, Yao
%A Zhao, Jun
%A Liu, Kang
%Y Li, Yunyao
%Y Rehm, Georg
%Y Tu, Mei
%A Yiwu
%S Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026)
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, USA
%@ 979-8-89176-394-4
%F yuan-etal-2026-r3a
%X Retrieval-augmented generation (RAG) plays a critical role in user-generated content (UGC) platforms, but its effectiveness critically depends on accurate query–document relevance assessment. Despite recent advances in applying large language models (LLMs) to relevance modeling, UGC platforms present unique challenges: 1) ambiguous user intent due to sparse user feedback in RAG scenarios, and 2) asymmetric relevance, where relevance is driven by localized answer-bearing content rather than global query–document similarity. To address these issues, we propose the Reinforced Reasoning model for Relevance Assessment (R³A), which decomposes relevance assessment into intent inference and evidence grounding. R³A leverages auxiliary high-clicked documents to infer latent query intent, and extracts verbatim evidence fragments to ground relevance decisions, reducing noise sensitivity and improving asymmetric relevance modeling. Experimental results demonstrate that R³A substantially outperforms strong baselines on offline benchmarks, while the distilled R³A-1.5B model achieves significant gains in large-scale online A/B testing, effectively balancing performance and practical deployability.
%U https://aclanthology.org/2026.acl-industry.10/
%P 127-140
Markdown (Informal)
[R³A: Reinforced Reasoning for Relevance Assessment for RAG in User-Generated Content Platforms](https://aclanthology.org/2026.acl-industry.10/) (Yuan et al., ACL 2026)
ACL
- Xiaowei Yuan, Lei Jin, Haoxin Zhang, Ziyang Huang, Yan Gao, Yiwu, Yao Hu, Jun Zhao, and Kang Liu. 2026. R³A: Reinforced Reasoning for Relevance Assessment for RAG in User-Generated Content Platforms. In Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026), pages 127–140, San Diego, California, USA. Association for Computational Linguistics.