@inproceedings{li-etal-2026-kare,
title = "{KARE}-{RAG}: Knowledge-Aware Refinement and Enhancement for {RAG}",
author = "Li, Yongjian and
Chu, HaoCheng and
Yan, Yukun and
Liu, Zhenghao and
Yu, Shi and
Zeng, Zheni and
Wang, Ruobing and
Song, Sen and
Liu, Zhiyuan and
Sun, Maosong",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.2311/",
pages = "49601--49623",
ISBN = "979-8-89176-462-0",
abstract = "Retrieval-Augmented Generation (RAG) equips large language models with external knowledge and is central to knowledge-intensive tasks. As RAG systems enter real-world use, generators must reliably leverage retrieved evidence. Recent fine-tuning methods improve adaptation to RAG scenarios, but optimization remains challenging because retrieval may return incomplete, fragmented, noisy, or conflicting contexts. Complex tasks further require fine-grained evidence dependencies. These challenges make high-quality supervision costly and limit generalization.\par We present KARE-RAG (Knowledge-Aware Refinement and Enhancement for RAG), a training-time scaffolded alignment framework. It uses structured knowledge representations as temporary scaffolds to expose evidence organization, support localized factual refinement, and construct fine-grained preference pairs. In our main implementation, an expert LLM refines a lightweight graph-structured evidence sketch. The generator is optimized with token-weighted Dense Direct Preference Optimization (DDPO), which focuses learning on edited scaffold regions. Scaffolds are used only for data construction and training supervision. At inference time, the model runs as standard Vanilla RAG without graph construction, extra retrieval, or latency overhead. Experiments show that KARE improves transfer across the evaluated QA and relation extraction datasets with limited training data, while leaving general capabilities largely unchanged. KARE can also complement existing RAG training objectives as an additional alignment stage."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="li-etal-2026-kare">
<titleInfo>
<title>KARE-RAG: Knowledge-Aware Refinement and Enhancement for RAG</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yongjian</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">HaoCheng</namePart>
<namePart type="family">Chu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yukun</namePart>
<namePart type="family">Yan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhenghao</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shi</namePart>
<namePart type="family">Yu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zheni</namePart>
<namePart type="family">Zeng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ruobing</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sen</namePart>
<namePart type="family">Song</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhiyuan</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Maosong</namePart>
<namePart type="family">Sun</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-462-0</identifier>
</relatedItem>
<abstract>Retrieval-Augmented Generation (RAG) equips large language models with external knowledge and is central to knowledge-intensive tasks. As RAG systems enter real-world use, generators must reliably leverage retrieved evidence. Recent fine-tuning methods improve adaptation to RAG scenarios, but optimization remains challenging because retrieval may return incomplete, fragmented, noisy, or conflicting contexts. Complex tasks further require fine-grained evidence dependencies. These challenges make high-quality supervision costly and limit generalization.\par We present KARE-RAG (Knowledge-Aware Refinement and Enhancement for RAG), a training-time scaffolded alignment framework. It uses structured knowledge representations as temporary scaffolds to expose evidence organization, support localized factual refinement, and construct fine-grained preference pairs. In our main implementation, an expert LLM refines a lightweight graph-structured evidence sketch. The generator is optimized with token-weighted Dense Direct Preference Optimization (DDPO), which focuses learning on edited scaffold regions. Scaffolds are used only for data construction and training supervision. At inference time, the model runs as standard Vanilla RAG without graph construction, extra retrieval, or latency overhead. Experiments show that KARE improves transfer across the evaluated QA and relation extraction datasets with limited training data, while leaving general capabilities largely unchanged. KARE can also complement existing RAG training objectives as an additional alignment stage.</abstract>
<identifier type="citekey">li-etal-2026-kare</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.2311/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>49601</start>
<end>49623</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T KARE-RAG: Knowledge-Aware Refinement and Enhancement for RAG
%A Li, Yongjian
%A Chu, HaoCheng
%A Yan, Yukun
%A Liu, Zhenghao
%A Yu, Shi
%A Zeng, Zheni
%A Wang, Ruobing
%A Song, Sen
%A Liu, Zhiyuan
%A Sun, Maosong
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F li-etal-2026-kare
%X Retrieval-Augmented Generation (RAG) equips large language models with external knowledge and is central to knowledge-intensive tasks. As RAG systems enter real-world use, generators must reliably leverage retrieved evidence. Recent fine-tuning methods improve adaptation to RAG scenarios, but optimization remains challenging because retrieval may return incomplete, fragmented, noisy, or conflicting contexts. Complex tasks further require fine-grained evidence dependencies. These challenges make high-quality supervision costly and limit generalization.\par We present KARE-RAG (Knowledge-Aware Refinement and Enhancement for RAG), a training-time scaffolded alignment framework. It uses structured knowledge representations as temporary scaffolds to expose evidence organization, support localized factual refinement, and construct fine-grained preference pairs. In our main implementation, an expert LLM refines a lightweight graph-structured evidence sketch. The generator is optimized with token-weighted Dense Direct Preference Optimization (DDPO), which focuses learning on edited scaffold regions. Scaffolds are used only for data construction and training supervision. At inference time, the model runs as standard Vanilla RAG without graph construction, extra retrieval, or latency overhead. Experiments show that KARE improves transfer across the evaluated QA and relation extraction datasets with limited training data, while leaving general capabilities largely unchanged. KARE can also complement existing RAG training objectives as an additional alignment stage.
%U https://aclanthology.org/2026.findings-emnlp.2311/
%P 49601-49623
Markdown (Informal)
[KARE-RAG: Knowledge-Aware Refinement and Enhancement for RAG](https://aclanthology.org/2026.findings-emnlp.2311/) (Li et al., Findings 2026)
ACL
- Yongjian Li, HaoCheng Chu, Yukun Yan, Zhenghao Liu, Shi Yu, Zheni Zeng, Ruobing Wang, Sen Song, Zhiyuan Liu, and Maosong Sun. 2026. KARE-RAG: Knowledge-Aware Refinement and Enhancement for RAG. In Findings of the Association for Computational Linguistics: EMNLP 2026, pages 49601–49623, Budapest, Hungary. Association for Computational Linguistics.