计算机应用研究2025，Vol.42Issue(11)：3236-3242,7.DOI:10.19734/j.issn.1001-3695.2025.04.0103
面向知识图谱推理共享的区块链优化方法
Optimization method for blockchain oriented to knowledge graph reasoning sharing
摘要
Abstract
This paper studied the problem of insufficient knowledge completeness in the current blockchain-based knowledge sharing architecture and proposed a complete KG sharing method based on blockchain(CKGC).By integrating blockchain with KG and introducing a reasoning mechanism,it prompted various industries to construct standards based on KG and created more accurate KGs,thereby enhancing the efficiency of knowledge utilization.It introduced the blockchain oracle to facilitate the interaction of reasoning information between on-chain and off-chain KGs.Mean while,it designed a Schnorr signature algo-rithm based on a threshold mechanism to improve the signature verification efficiency of the oracle when uploading KG reason-ing results to the blockchain.It designed a KG reasoning method based on LLM to improve KG,leveraged the semantic under-standing ability of LLM to generate high-quality logical rules for KG reasoning and enhance the richness and accuracy of KG.Experimental results show that Schnorr threshold signature algorithm improves the signature efficiency by 60％compared to Schnorr aggregate signature algorithm.KG reasoning method based on LLM is evaluated on three large KGs,confirming its effectiveness and advancement.CKGC provides an efficient and trustworthy technical path for knowledge collaboration and sha-ring.关键词
区块链/知识共享/知识图谱/预言机/签名算法/大语言模型/知识图谱推理Key words
blockchain/knowledge sharing/knowledge graph(KG)/oracle/signature algorithm/large language model(LLM)/knowledge graph reasoning分类
计算机与自动化引用本文复制引用
许振彬,郭荣新,施一帆,黄凯..面向知识图谱推理共享的区块链优化方法[J].计算机应用研究,2025,42(11):3236-3242,7.基金项目
国家自然科学基金青年项目(62306122) （62306122）
福建省科技项目引导性项目(2023H0012) （2023H0012）
