EMNLP Findings2024
TRACE the Evidence: Constructing Knowledge-Grounded Reasoning Chains for Retrieval-Augmented Generation
Abstract
Retrieval-augmented generation (RAG) offers an effective approach for addressing question answering (QA) tasks. However, the imperfections of the retrievers in RAG models often result in the retrieval of irrelevant information, which could introduce noise and degrade the performance, especially when handling multihop questions that require multiple steps of reasoning. To enhance the multi-hop reasoning ability of RAG models, we propose TRACE 1 . TRACE constructs knowledge-grounded reasoning chains, which are a series of logically connected knowledge triples, to identify and integrate supporting evidence from the retrieved documents for answering questions. Specifically, TRACE employs a KG Generator to create a knowledge graph (KG) from the retrieved documents, and then uses a novel Autoregressive Reasoning Chain Constructor to build reasoning chains. Experimental results on three multi-hop QA datasets show that TRACE achieves an average performance improvement of up to 14.03% compared to using all the retrieved documents. Moreover, the results indicate that using reasoning chains as context, rather than the entire documents, is often sufficient to correctly answer questions.
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
10 top-tier papers have cited it since. Know what they found before your reviewers do.
Lune can draw on
the full paper · the top-tier papers that cite it · related work
Free to start. No credit card required.
Lineage
- NeocorRAG: Less Irrelevant Information, More Explicit Evidence, and More Effective Recall via Evidence Chains cites Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
- CompactRAG: Reducing LLM Calls and Token Overhead in Multi-Hop Question Answering cites Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
- One Token Can Help! Learning Scalable and Pluggable Virtual Tokens for Retrieval-Augmented Large Language Models cites RA-DIT: Retrieval-Augmented Dual Instruction Tuning
- One Token Can Help! Learning Scalable and Pluggable Virtual Tokens for Retrieval-Augmented Large Language Models cites Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
- RA-DIT: Retrieval-Augmented Dual Instruction Tuning cites Large Language Models Can Be Easily Distracted by Irrelevant Context
- RA-DIT: Retrieval-Augmented Dual Instruction Tuning cites Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
Cited by top-tier papers10
- One Token Can Help! Learning Scalable and Pluggable Virtual Tokens for Retrieval-Augmented Large Language ModelsYutao Zhu, Zhaoheng Huang, Zhicheng Dou, Ji-Rong WenAAAI 2025 · 9 citations
- Demystifying Deep Search: A Holistic Evaluation with Hint-free Multi-Hop Questions and Factorised MetricsMaojia Song, Renhang Liu, Xinyu Wang, Yong Jiang et al.ICLR 2026 · 7 citations
- Cooperative Retrieval-Augmented Generation for Question Answering: Mutual Information Exchange and Ranking by Contrasting LayersYoumin Ko, Sungjong Seo, Hyunjoon KimNeurIPS 2025 · 2 citations
- NeocorRAG: Less Irrelevant Information, More Explicit Evidence, and More Effective Recall via Evidence ChainsShiyao Peng, Qianhe Zheng, Zhuodi Hao, Zichen Tang et al.WWW 2026
- CompactRAG: Reducing LLM Calls and Token Overhead in Multi-Hop Question AnsweringHao Yang, Zhiyu Yang, Xupeng Zhang, Wei Wei et al.WWW 2026
Builds on18
- Retrieval-Augmented Generation for Knowledge-Intensive NLP TasksPatrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni et al.NeurIPS 2020 · 19,162 citations
- Large Language Models Can Be Easily Distracted by Irrelevant ContextFreda Shi, Xinyun Chen, Kanishka Misra, Nathan Scales et al.ICML 2023 · 970 citations
- Reasoning on Graphs: Faithful and Interpretable Large Language Model ReasoningLinhao Luo, Yuan-Fang Li, Gholamreza Haffari, Shirui PanICLR 2024 · 499 citations
- Document Ranking with a Pretrained Sequence-to-Sequence ModelRodrigo Nogueira, Zhiying Jiang, Ronak Pradeep, Jimmy LinEMNLP Findings 2020 · 469 citations
- RA-DIT: Retrieval-Augmented Dual Instruction TuningXi Victoria Lin, Xilun Chen, Mingda Chen, Weijia Shi et al.ICLR 2024 · 229 citations
Related papers
- KiRAG: Knowledge-Driven Iterative Retriever for Enhancing Retrieval-Augmented GenerationJinyuan Fang, Zaiqiao Meng, Craig MacDonaldACL 2025
- KnowTrace: Bootstrapping Iterative Retrieval-Augmented Generation with Structured Knowledge TracingRui Li, Quanyu Dai, Zeyu Zhang, Xu Chen et al.KDD 2025 · 1 citation
- DTKG: Dual-Track Knowledge Graph-Verified Reasoning Framework for Multi-Hop QAChanghao Wang, Yanfang Liu, Xinxin Fan, Ao Tian et al.ICML 2026
- HopRAG: Multi-Hop Reasoning for Logic-Aware Retrieval-Augmented GenerationHao Liu, Zhengren Wang, Xi Chen, Zhiyu Li et al.ACL Findings 2025 · 24 citations
- Chain-of-Retrieval Augmented GenerationLiang Wang, Haonan Chen, Nan Yang, Xiaolong Huang et al.NeurIPS 2025 · 59 citations
