基于RDF与概率推理的不确定性知识表示算法

Uncertainty Knowledge Representation Algorithm Based on RDF and Probabilistic Reasoning

摘要

针对不确定性知识具有模糊性、随机性或不完整性,很难用单一方式准确表示和处理的问题,提出一种基于资源描述框架(RDF)与概率推理的不确定性知识表示算法.首先,采用RDF图描述不确定性知识数据样本,构建不确定性知识元语句及其层次关系,进而得到不确定性知识RDF图模式与标准语句模式;其次,用模糊Petri网表示不确定性知识,定义模糊Petri网八元组,采用概率软逻辑推理方式构建模糊推理规则,并对逻辑推理规则进行约束;最后,通过特定算子推理,在库可信度数值稳定时输出不确定性知识表示结果.实验结果表明:该方法构建的标准语句语义丰富度数值均高于0.8;在逻辑规则增加至220个时,出现逻辑矛盾的次数仅为3,概率为1.36%;不同不确定性知识表示的确定度均高于0.9,表明算法在表示不确定性知识时精确性高,能有效捕捉并描述知识内部变量的逻辑关系.

Abstract

Aiming at the problem that it was difficult to accurately represent and process uncertain knowledge in a single way due to its ambiguity, randomness or incompleteness, we proposed an uncertainty knowledge representation algorithm based on resource description framework (RDF) and probabilistic reasoning. Firstly, RDF graph was used to describe uncertain knowledge data samples, and uncertain knowledge metastatements and their hierarchical relationships were constructed to obtain uncertain knowledge RDF graph patterns and standard statement patterns. Secondly, fuzzy Petri nets were used to represent uncertain knowledge and define fuzzy Petri net octets. The probabilistic soft logic reasoning was used to construct fuzzy inference rules and constrain the logical inference rules. Finally, through specific operator inference, we output uncertain knowledge representation results when the library credibility value was stable. The experimental results show that the semantic richness values of the standard sentences constructed by proposed method are all higher than 0.8. When the number of logical rules increases to 220, there are only 3 occurrences of logical contradictions, with a probability of 1.36%. The certainty of different uncertain knowledge representations is all above 0.9, indicating that the algorithm has high accuracy in representing uncertain knowledge and can effectively capture and describe the logical relationships of variables within the knowledge.

关键词

资源描述框架图 / 概率软推理 / 不确定性 / 知识表示 / 模糊Petri网

Key words

resource description framework graph / probabilistic soft reasoning / uncertainty / knowledge representation / fuzzy Petri net

参考文献

[1] 陈冠恒, 郭子瑜, 梅广旭, 等. 一种针对关系不确定性的贝叶斯异质图神经网络[J]. 计算机学报, 2023, 46(3):552-567.

[2] (CHEN G H, GUO Z Y, MEI G X, et al. A Bayesian Heterogeneous Graph Neural Network for Relational Uncertainty[J]. Chinese Journal of Computers, 2023, 46(3):552-567.)

[3] 王帆, 韩忠义, 尹义龙. 伪标签不确定性估计的源域无关鲁棒域自适应[J]. 软件学报, 2022, 33(4):1183-1199.

[4] (WANG F, HAN Z Y, YIN Y L. Source Free Robust Domain Adaptation Based on Pseudo Label Uncertainty Estimation[J]. Journal of Software, 2022, 33(4):1183-1199.)

[5] 姚建军, 李剑宇, 岳昆, 等. 基于概率推理的知识图谱链接预测方法[J]. 计算机集成制造系统, 2023, 29(10):3483-3495.

[6] (YAO J J, LI J Y, YUE K, et al. Approach for Link Prediction of Knowledge Graph Based on Probabilistic Inferences[J]. Computer Integrated Manufacturing Systems, 2023, 29(10):3483-3495.)

[7] 谭开成, 罗继亮, 林鑫杰, 等. 基于知识Petri网的确定性和不确定性联合推理[J]. 控制理论与应用, 2023, 40(3):531-539.

[8] (TAN K C, LUO J L, LIN X J, et al. Certainty and Uncertainty Joint Reasoning Based on Knowledge Petri Nets[J]. Control Theory & Applications, 2023, 40(3):531-539.)

[9] 徐遥, 何世柱, 刘康, 等. 基于Beta分布和半监督学习的非确定性知识图谱嵌入模型[J]. 中文信息学报, 2022, 36(10):54-62.

[10] (XU Y, HE S Z, LIU K, et al. Uncertain Knowledge Graph Embedding by Beta Distribution and Semi-supervised Learning[J]. Journal of Chinese Information Processing, 2022, 36(10):54-62.)

[11] 李健京, 李贯峰, 秦飞舟, 等. 基于不确定知识图谱嵌入的多关系近似推理模型[J]. 计算机应用, 2024, 44(6):1751-1759.

[12] (LI J J, LI G F, QIN F Z, et al. Multi-relation Approximate Reasoning Model Based on Uncertain Knowledge Graph Embedding[J]. Journal of Computer Applications, 2024, 44(6):1751-1759.)

[13] 王翔. 基于滑动窗口的流式RDF数据的模式匹配方法[J]. 计算机工程与设计, 2024, 45(5):1458-1464.

[14] (WANG X. Pattern Matching for Streaming RDF Graph over Sliding Windows[J]. Computer Engineering and Design, 2024, 45(5):1458-1464.)

[15] 冯辉, 何伊竞, 徐海祥, 等. 基于模糊推理的STPF-AIMM水面目标跟踪算法[J]. 华中科技大学学报(自然科学版), 2023, 51(8):109-114.

[16] (FENG H, HE Y J, XU H X, et al. Water Surface Target Tracking Algorithm Based on Fuzzy Inference STPF-AIMM[J]. Journal of Huazhong University of Science and Technology(Natural Science Edition), 2023, 51(8):109-114.)

[17] 李彦柯, 祁志卫, 李剑宇, 等. 知识图谱引导的贝叶斯网概率推理[J]. 小型微型计算机系统, 2022, 43(6):1272-1277.

[18] (LI Y K, QI Z W, LI J Y, et al. Knowledge Graph Induced Probabilistic Inference of Bayesian Network[J]. Journal of Chinese Computer Systems, 2022, 43(6):1272-1277.)

[19] 曾繁慧, 胡光闪, 孙慧, 等. 因素空间理论下的因果概率推理分类算法研究[J]. 智能系统学报, 2024, 19(4):1042-1051.

[20] (ZENG F H, HU G S, SUN H, et al. A Causal Probabilistic Inference Classification Algorithm Based on Factor Space Theory[J]. CAAI Transactions on Intelligent Systems, 2024, 19(4):1042-1051.)

[21] 李雷, 徐浩, 吴素萍. 基于DDPG的三维重建模糊概率点推理[J]. 自动化学报, 2022, 48(4):1105-1118.

[22] (LI L, XU H, WU S P. Fuzzy Probability Points Reasoning for 3D Reconstruction via Deep Deterministic Policy Gradient[J]. Acta Automatica Sinica, 2022, 48(4):1105-1118.)

[23] 张莉丽, 王睿伯, 王晓东, 等. 自治故障管理系统推理规则的智能学习技术[J]. 计算机工程与科学, 2023, 45(5):772-781.

[24] (ZHANG L L, WANG R B, WANG X D, et al. Inference Rule Learning in Autonomous Fault Management Systems[J]. Computer Engineering & Science, 2023, 45(5):772-781.)

[25] 汤宇轩, 申彦明, 齐恒, 等. 基于同构查询校准的知识图谱复杂逻辑推理[J]. 中文信息学报, 2024, 38(9):24-35.

[26] (TANG Y X, SHEN Y M, QI H, et al. Complex Logical Reasoning in Knowledge Graph with Homogeneous Query Revisor[J]. Journal of Chinese Information Processing, 2024, 38(9):24-35.)

[27] 罗敏霞, 徐东辉. 区间值模糊推理的逻辑度量空间[J]. 智能系统学报, 2023, 18(3):613-618.

[28] (LUO M X, XU D H. Logical Metric Spaces for Interval-Valued Fuzzy Reasoning[J]. CAAI Transactions on Intelligent Systems, 2023, 18(3):613-618.)

[29] 吴立波, 黄玉芳. 基于DNA链置换的逻辑推理问题研究[J]. 计算机科学, 2022, 49(1):259-263.

[30] (WU L B, HUANG Y F. Logical Reasoning Based on DNA Strand Displacement[J]. Computer Science, 2022, 49(1):259-263.)

基金资助

宁夏自然科学基金(2024AAC03214)