Repair Before Reinforce: Context-Augmented Knowledge Graph Reasoning for Multi-Hop Question Answering

• URL: http://arxiv.org/abs/2609.12230v1
• Date: Thu, 10 Sep 2026 21:39:11 GMT
• Title: Repair Before Reinforce: Context-Augmented Knowledge Graph Reasoning for Multi-Hop Question Answering
• Abstract summary: We propose a context-augmented training framework for multi-hop question-answering.<n>Although generally applicable, we validate the framework in the context of disease-specific knowledge graphs.<n>Across both diseases, context-augmented supervision consistently improves multi-hop performance over KG-only supervision.
• Score: 4.464939140209426
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Question-answering often requires reasoning across multiple connected facts rather than retrieving a single isolated relation. Knowledge graphs (KGs) provide a structured way to represent such facts, but training large language models (LLMs) only on isolated KG head-relation-tail triples may limit their ability to learn the surrounding context needed for multi-hop reasoning. In this work, we propose a context-augmented training framework for multi-hop question-answering. Although generally applicable, we validate the framework in the context of disease-specific KGs, extracted using a reliable KG extraction framework called GraphMERT, for Gastroparesis and Diabetes. For each primary KG triple, we attach supporting triples extracted from the same source text chunk to form a context graph (CG). This creates two supervision settings: KG-grounded supervision, which uses only the target KG triple or path, and CG-grounded supervision, which uses the target KG triple or path together with supporting context triples. We train the Qwen3-14B model using supervised fine-tuning (SFT) under both settings, producing KGModel and CGModel variants. To strengthen the lower-hop factual foundation of the models, we introduce an LLM-judged, history-aware adaptive repair pipeline that identifies unresolved one-hop failures, continually fine-tunes on targeted repair examples, and removes or quarantines problematic noisy triples. This repair stage enables the models to reach 100% accuracy on the cleaned retained one-hop validation sets. Finally, we employ reinforcement learning (RL) using lower-hop question-answer items and evaluate generalization on harder 3-hop, 4-hop, and 5-hop tasks. Across both diseases, context-augmented supervision consistently improves multi-hop performance over KG-only supervision. RL initialized from repaired SFT checkpoints yields larger and more stable gains.

Related papers

• Travel-Oriented Reasoning Large Language Model via Domain-Specific Knowledge Graphs [0.0]
  Large language models (LLMs) demonstrate broad reasoning abilities but struggle with accuracy and reliability in specialized domains such as travel.<n>We propose a modular pipeline for building a travel-domain reasoning LLM grounded in an expert-designed knowledge graph.<n>Our pipeline integrates a travel KG that encodes domain entities and their relationships, a supervised fine-tuning stage that embeds the domain knowledge into a reasoning-capable LLM.
  arXiv  Detail & Related papers  (2026-06-28T07:53:11Z)
• KG-Hopper: Empowering Compact Open LLMs with Knowledge Graph Reasoning via Reinforcement Learning [9.329800077564673]
  Large Language Models (LLMs) demonstrate impressive natural language capabilities but often struggle with knowledge-intensive reasoning tasks.<n>We propose KG-Hopper, a novel Reinforcement Learning framework that empowers compact open LLMs with the ability to perform integrated multi-hop KG reasoning.<n>We show that KG-Hopper consistently outperforms larger multi-step systems (up to 70B) and achieves competitive performance with proprietary models.
  arXiv  Detail & Related papers  (2026-03-22T23:07:29Z)
• Mitigating KG Quality Issues: A Robust Multi-Hop GraphRAG Retrieval Framework [11.158010513386664]
  We propose C2RAG (Constraint-Checked Retrieval-Augmented Generation), a framework aimed at robust multi-hop retrieval over the imperfect KG.<n>C2RAG consistently outperforms the latest baselines by 3.4% EM and 3.9% F1 on average.
  arXiv  Detail & Related papers  (2026-03-16T05:08:41Z)
• DPRM: A Dual Implicit Process Reward Model in Multi-Hop Question Answering [62.11407895095892]
  In question answering tasks, Chain of Thought (CoT) guides large language models through multi-step reasoning, and Knowledge Graphs (KGs) reduce hallucinations via semantic matching.<n>Traditional Process Reward Models (PRMs) evaluate the reasoning process but require costly human annotations or rollout generation.<n>We propose the DPRM (Dual Implicit Process Reward Model), which trains two implicit PRMs for CoT and KG reasoning in MHQA tasks.
  arXiv  Detail & Related papers  (2025-11-11T15:41:22Z)
• Decoding on Graphs: Faithful and Sound Reasoning on Knowledge Graphs through Generation of Well-Formed Chains [66.55612528039894]
  Knowledge Graphs (KGs) can serve as reliable knowledge sources for question answering (QA) We present DoG (Decoding on Graphs), a novel framework that facilitates a deep synergy between LLMs and KGs. Experiments across various KGQA tasks with different background KGs demonstrate that DoG achieves superior and robust performance.
  arXiv  Detail & Related papers  (2024-10-24T04:01:40Z)
• UniKGQA: Unified Retrieval and Reasoning for Solving Multi-hop Question Answering Over Knowledge Graph [89.98762327725112]
  Multi-hop Question Answering over Knowledge Graph(KGQA) aims to find the answer entities that are multiple hops away from the topic entities mentioned in a natural language question. We propose UniKGQA, a novel approach for multi-hop KGQA task, by unifying retrieval and reasoning in both model architecture and parameter learning.
  arXiv  Detail & Related papers  (2022-12-02T04:08:09Z)
• Contrastive Knowledge Graph Error Detection [11.637359888052014]
  We propose a novel framework - ContrAstive knowledge Graph Error Detection (CAGED) CAGED introduces contrastive learning into KG learning and provides a novel way of modeling KG. It outperforms state-of-the-art methods in KG error detection.
  arXiv  Detail & Related papers  (2022-11-18T05:01:19Z)
• Few-shot Relational Reasoning via Connection Subgraph Pretraining [81.30830261527231]
  Connection Subgraph Reasoner (CSR) can make predictions for the target few-shot task directly without the need for pre-training. Our framework can already perform competitively to existing methods on target few-shot tasks.
  arXiv  Detail & Related papers  (2022-10-13T04:35:14Z)
• Collaborative Knowledge Graph Fusion by Exploiting the Open Corpus [59.20235923987045]
  It is challenging to enrich a Knowledge Graph with newly harvested triples while maintaining the quality of the knowledge representation. This paper proposes a system to refine a KG using information harvested from an additional corpus.
  arXiv  Detail & Related papers  (2022-06-15T12:16:10Z)
• SMORE: Knowledge Graph Completion and Multi-hop Reasoning in Massive Knowledge Graphs [147.73127662757335]
  We present scalable Multi-hOp REasoning (SMORE), the first general framework for both single-hop and multi-hop reasoning in Knowledge Graphs (KGs) Using a single machine SMORE can perform multi-hop reasoning in Freebase KG (86M entities, 338M edges), which is 1,500x larger than previously considered KGs. SMORE increases throughput (i.e., training speed) over prior multi-hop KG frameworks by 2.2x with minimal GPU memory requirements.
  arXiv  Detail & Related papers  (2021-10-28T05:02:33Z)
• Improving Embedded Knowledge Graph Multi-hop Question Answering by introducing Relational Chain Reasoning [8.05076085499457]
  Knowledge Base Question Answer (KBQA) to answer userquestions from a knowledge base (KB) by identifying reasoning between topic entity and answer. As a complex branchtask of KBQA, multi-hop KGQA requires reasoning over multi-hop relational chains preserved in structured KG.
  arXiv  Detail & Related papers  (2021-10-25T06:53:02Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.