EMNLP Findings2025
SCE: Semantic Consistency Enhanced Reinforcement Learning for Multi-Hop Knowledge Graph Reasoning
Yanwen Huang, Yao Liu, Qiao Liu, Rui Hou, Tingting Dai
Abstract
Multi-hop reasoning with reinforcement learning has proven effective in discovering inference paths in incomplete knowledge graphs. However, a major challenge remains: spurious paths (incorrect reasoning paths that accidentally lead to correct answers) often arise due to reward mechanisms that prioritize final results over reasoning quality. While existing approaches attempt to mitigate this issue using external rules, they often neglect the internal semantic consistency between the target triple and the intermediate triples along the reasoning path. In this paper, we propose a novel framework, Semantic Consistency Enhanced Reinforcement Learning (SCE), which incorporates semantic consistency into the reward function to guide multi-hop reasoning. Experimental results demonstrate that SCE outperforms strong baseline methods and facilitates the discovery of more interpretable reasoning paths. 1
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.
Lineage
- ANet: A Scalable Path-based Reasoning Approach for Knowledge Graphs cites Reasoning on Knowledge Graphs with Debate Dynamics
- ANet: A Scalable Path-based Reasoning Approach for Knowledge Graphs cites Learning to Walk with Dual Agents for Knowledge Graph Reasoning
- Learning to Walk with Dual Agents for Knowledge Graph Reasoning cites Reasoning on Knowledge Graphs with Debate Dynamics
Builds on7
- A*Net: A Scalable Path-based Reasoning Approach for Knowledge GraphsZhaocheng Zhu, Xinyu Yuan, Michael Galkin, Louis-Pascal A. C. Xhonneux et al.NeurIPS 2023 · 103 citations
- Learning to Walk with Dual Agents for Knowledge Graph ReasoningDenghui Zhang, Zixuan Yuan, Hao Liu, Xiaodong Lin et al.AAAI 2022 · 61 citations
- Reasoning on Knowledge Graphs with Debate DynamicsMarcel Hildebrandt, Jorge Andres Quintero Serna, Yunpu Ma, Martin Ringsquandl et al.AAAI 2020 · 59 citations
- GaussianPath: A Bayesian Multi-Hop Reasoning Framework for Knowledge Graph ReasoningGuojia Wan, Bo DuAAAI 2021 · 59 citations
- RLogic: Recursive Logical Rule Learning from Knowledge GraphsKewei Cheng, Jiahao Liu, Wei Wang, Yizhou SunKDD 2022 · 57 citations
Related papers
- Rule-Aware Reinforcement Learning for Knowledge Graph ReasoningZhongni Hou, Xiaolong Jin, Zixuan Li, Long BaiACL Findings 2021 · 35 citations
- SQUIRE: A Sequence-to-sequence Framework for Multi-hop Knowledge Graph ReasoningYushi Bai, Xin Lv, Juanzi Li, Lei Hou et al.EMNLP 2022 · 19 citations
- Multi-level Recommendation Reasoning over Knowledge Graphs with Reinforcement LearningXiting Wang, Kunpeng Liu, Dongjie Wang, Le Wu et al.WWW 2022 · 125 citations
- DREAM: Adaptive Reinforcement Learning based on Attention Mechanism for Temporal Knowledge Graph ReasoningShangfei Zheng, Hongzhi Yin, Tong Chen, Quoc Viet Hung Nguyen et al.SIGIR 2023 · 27 citations
- MCMH: Learning Multi-Chain Multi-Hop Rules for Knowledge Graph ReasoningLu Zhang, Mo Yu, Tian Gao, Yue YuEMNLP Findings 2020 · 2 citations
