The current construction of knowledge graphs in the petroleum industry faces issues of semantic heterogeneity and weak interoperability. As the schema layer of the knowledge graph, the complete and standardized ontology is the key to ensuring the quality of knowledge graphs. This paper first focuses on the semantic constraints added in the latest version of the MDR standard, expands and defines the mapping rules from MDR to OWL ontology, and establishes the bridge between metadata and ontology semantics. Secondly, this paper conducts an investigation and research on the relevant data models and standards in the petroleum field and performs semantic analysis on the relevant data elements from different perspectives. On this basis, an exploration and development metadata registration platform based on MDR was developed to achieve standardized registration and management of domain metadata and related semantics. Finally, based on the proposed mapping rules, the construction of the domain ontology is completed. This paper provides a reference for semantic management and ontology construction based on metadata in vertical domains, thereby facilitating the standardized construction of rich semantic knowledge graphs.

Knowledge graphs (KGs) are graphical knowledge bases widely used to represent and organize knowledge, playing a crucial role in various artificial intelligence applications. The quality of a knowledge graph significantly impacts its utility. Knowledge graph completion (KGC) is a critical approach to enhancing the quality of knowledge graphs, typically categorized into structure-based methods and methods leveraging descriptive information. However, structure-based methods often suffer from data sparsity issues due to their reliance solely on the triadic structure of the knowledge graph. On the other hand, methods utilizing descriptive information face challenges in effectively leveraging and modeling such information. To address these issues, this study proposes AutoTransW, a novel model based on asymmetric translation and automatic representation of entity types. AutoTransW embeds entities and relations into a complex space, treating relations as rotational operations from a head entity to a tail entity. Entity-specific triple encoders with hyperplane projection strategies are employed to embed entities and relations into the complex space. Secondly, type-specific triple encoders learn type embeddings through a relation-aware projection mechanism. Additionally, the incorporation of neighboring information about entities also enhances the model’s learning capabilities. Experimental results on the WN11, WN18, FB13, and FB15K datasets demonstrate that AutoTransW outperforms current state-of-the-art baselines in link prediction and triple classification tasks, confirming the model’s effectiveness.

The Electronics Editorial Office retracts the article “A Global Structural Hypergraph Convolutional Model for Bundle Recommendation” [...]

In recent years, emerging technologies such as immersive virtual reality (IVR) and Metaverse are pointing to new directions for the future of education. To summarise the current research status and development trend of immersive virtual reality in the field of education, this paper used the CiteSpace tool to search the SSCI literature included in the Web of Science under the theme of “Immersive virtual reality” or “Educational technology”. The results show a total of 1293 SSCI publications between 1996 and 2022, with the number of studies in this area increasing year on year. Current research hot spots focus on the application of IVR technology, evaluation and effectiveness research, and curriculum design and teaching. Analysis of node network diagrams and knowledge graphs suggest that future trends in the field will be to deepen immersive experiences, increase interactivity, and continue to explore more pedagogical value and application scenarios. In addition, there are increasingly close partnerships between research institutions and large-scale collaborative research is taking place internationally. In future research, researchers can explore scenarios for the application of Metaverse and the design of assessment systems that can facilitate the development of immersive education.

传统推荐算法缺少对图结构的隐含信息及上下文信息的利用,从而可能降低推荐效果.为提高传统推荐算法的推荐效果,提出基于图神经网络的推荐模型.该模型基于图的高阶连通性理论,使用图神经网络挖掘用户.物品二部图中的隐含信息,并由一阶扩展到多阶,从而获取更精确的嵌入式表示和推荐效果;在更新过程中考虑上下文信息,有利于理解上下文间的交互关系.并将该模型在Yelp-OS、Yelp-NC和Amazon-book数据集上进行实验,实验结果表明,在HR(Hit Ratio)和NDCG(Normalized Discounted Cumulative Gain)指标上均优于相关对比算法,证明该算法可优化推荐效果,提升推荐质量.

Bundle recommendations provide personalized suggestions to users by combining related items into bundles, aiming to enhance users’ shopping experiences and boost merchants’ sales revenue. Existing solutions based on graph neural networks (GNN) face several significant challenges: (1) it is demanding to explicitly model multiple complex associations using standard graph neural networks, (2) numerous additional nodes and edges are introduced to approximate higher-order associations, and (3) the user–bundle historical interaction data are highly sparse. In this work, we propose a global structural hypergraph convolutional model for bundle recommendation (SHCBR) to address the above problems. Specifically, we jointly incorporate multiple complex interactions between users, items, and bundles into a relational hypergraph without introducing additional nodes and edges. The hypergraph structure inherently incorporates higher-order associations, thereby alleviating the training burden on neural networks and the dilemma of scarce data effectively. In addition, we design a special matrix propagation rule that captures non-pairwise complex relationships between entities. Using item nodes as links, structural hypergraph convolutional networks learn representations of users and bundles on a relational hypergraph. Experiments conducted on two real-world datasets demonstrate that the SHCBR outperforms the state-of-the-art baselines by 11.07–25.66% on Recall and 16.81–33.53% on NDCG. Experimental results further indicate that the approach based on hypergraphs can offer new insights for addressing bundle recommendation challenges. The codes and datasets have been publicly released on GitHub.

Abstract: At present, procedural knowledge is mainly represented by production rules. Although this representation is concise, problems exist on lack of semantic elements and single semantic relationship, which is not conducive to the representation of procedural knowledge and hinders the teaching and application of procedural knowledge. Therefore, this paper firstly analyzes the semantic elements that procedural knowledge should have, and the lack of time and space elements in traditional representation methods. Combining with the cognitive background of procedural knowledge and introducing event cognition, the representation method of procedural knowledge graph and the construction process model are obtained. The verification proves that the event graph is a procedural knowledge representation method with complete semantic roles and rich semantic relations.

近年来,随着中国的经济和人们生活水平的迅速提高,对能源开采和运输的需求日益增加,油气运输管道作为重要的能源运输方式之一,其总长度有了巨大的增长.而管道泄漏检测技术是保证油气管道高效、稳定运行的关键技术之一.为此,首先对管道泄漏检测以及定位技术的发展历程和现有研究成果进行了总结,并对各种常见方法进行了归类.其次研究了多种常见管道泄漏检测方法的基本原理和操作方法;然后根据实际工业情况阐述了各个检测方法的特点和适用应用场景;最后,对管道泄漏检测技术存在的挑战和未来研究的重心进行分析和展望.

程序性知识在人类学习与认知过程中扮演了重要的角色,目前对程序性知识的组织都是通过产生式系统实现,是一种非语义化、无法体现知识间的关系且不易维护的方法.为此,将ACT-R(Adaptive Control of Thought-Rational)理论、事件认知理论与文献中的事件模型相融合,提出一种能表征程序性知识的事件七元组语义模型.在此基础上,以Java语言中的程序性知识为例,构建了程序性知识事件本体模型,并设计了基于程序性知识事件的ACT-R认知模型对其进行验证.实验结果表明,程序性知识事件本体能实现对程序性知识及其之间关系语义完整性描述,为程序性知识教学资源组织与基于程序性知识的网络教学等提供了新的思路.

依托新工科背景,针对目前非计算机专业程序设计课程中存在的问题,提出利用思维导图实现程序设计认知结构建构,促进学生新知识的内化、扩展以及整体性构建,培养学生的计算思维,从而提高学生的程序设计应用和创新能力.

后疫情时期,高校教学形式进入线上+线下融合的时代,融合时代如何促进学习者深度学习发生值得深度研究.分析了线上+线下融合教学环境中深度学习发生的必要性,探讨了教育大数据在促进融合环境中学习者深度学习发展的优势,构建数据驱动的深度学习教学模式,以"传播学"课程为例叙述了该模式的教学实践过程,并对该模式的实践效果进行了说明.

现有地质元数据标准普遍表现为缺乏语义、比较零散、体系不够完善等缺点.为解决地质知识标准的组织,在对国内外地质领域内标准以及知识图谱研究的基础上,以地层知识和石油地质知识为依据构建了地质知识标准化模型,分别以地层和石油地质为应用背景,采用知识图谱技术对他们进行了实现验证.结果 表明,所提出的标准化能为地质知识组织的标准化提供标准支撑.

[目的/意义]在人工智能领域,知识图谱作为一种新的知识表示方式备受青睐,将其作为数据支撑的应用程序层出不穷.但是当前针对知识图谱的相关研究主要集中在知识图谱构建技术和应用两方面,而知识图谱质量评估仍处于萌芽阶段,特别是缺少与用户质量需求相关的数据质量模型研究.因此构建知识图谱质量模型对于知识图谱质量评估以及开发高质量应用具有重要意义.[方法/过程]文章通过梳理目前知识图谱各应用领域的业务质量需求,将其与质量维度进行映射,并基于W3C的DQV词表规范,遵循最小化本体原则和重用现有词表原则,采用本体栈的方式,构建出一个可扩展、健壮性强的知识图谱质量模型.[结果/结论]文章提出的质量模型提供了完备的、统一的、规范的术语体系来描述知识图谱质量的各个要素,以此来帮助用户检索符合自身业务需求的知识图谱,达到知识发现的目标.

为解决中文数据的重复检测问题,在SNM(Sorted-Neighborhood Method)算法的基础上提出了一种融合《同义词词林扩展版》和中文分词于一体的重复记录检测算法,通过利用《同义词词林扩展版》和Jaccard算法计算词语相似度,同时采用Python中Jieba中文分词的方式对语句进行分词,从而达到优化余弦相似度算法计算语句相似度.改进后的算法能有效的对记录为中文的字段和语句进行重复记录检测.并通过对某辅导机构学生测验信息数据集进行实验,实验结果表明,该算法的查全率及查准率比传统SNM算法有了很大提高.

本体融合已经成为学科领域知识图谱的重建和知识共享的重要途径,为解决本体融合领域缺乏标准融合框架问题,提出一种基于国际标准MFI4OR的本体融合模型.该模型提供一个标准的本体信息划分标准,即本体-本体构件-本体原子构件,实现对本体信息的管理和映射.在融合计算过程中,选择编辑距离算法并引入外部资源WordNet词典进行相似度计算.最后,以省级基金项目中的学习者模型构建需求为应用背景,将FOAF(Friend-Of-A-Friend)本体和RELATIONSHIP本体进行融合试验.结果 表明,该标准化融合模型能实现通用本体融合.

目前现有问答系统模型大多数都采用模板匹配的方式进行推理,对问题推理不够充分,因此,提出基于认知图谱的问答系统推理模型.依据专业领域知识作为知识源构建本体;并基于该认知图谱构建了"问题-关系"一对一的认知图谱问答系统模型.最后通过将问答问题划分为简单问题与复杂问题分别对问题进行处理,其中简单问题运用BERT+ CRF(Bidirectional Encoder Representations from Transformers+Conditional Random Field)模型进行模板匹配;针对复杂问题运用Node2vec生成子图后用GCN(Graph Convolutional Network)推理模型进行推理,将得出的答案作为输出结果.最后对所提出的模型通过井下作业领域进行了实验,结果表明认知图谱问答模型优于其他算法模型.

为能高效、准确地对关联数据的链接有效性进行评估,本文对国内外有关链接有效性的评估方法和技术进行了研究,发现目前对链接有效性研究的成果有些文献只有简单提及,成果很少.因此,本文提出了用于统一资源标次符(URI)有效性评估的?算法,通过理论分析证明了该算法的有效性和高效性.最后,利用DBpedia发布的开放数据进行实验验证,实验结果表明,该方法评估链接有效性提高了0.1％,评估效率为常规方法的4倍,验证了本文算法的准确性与高效性.

[目的/意义]随着知识图谱在互联网领域中的成功应用,使得知识图谱的应用迅速向垂直领域传播与扩散,越来越得到垂直领域的重视.我国各个领域经历了40多年的信息化建设,导致了一大批阻碍信息共享的“信息孤岛”,其根本原因是领域没有给予标准化足够的重视.无独有偶,40年后的今天,领域又一次无法回避知识表示中的标准化难题!目前知识图谱构建主要关注点在于知识图谱的构建技术,却很少考虑知识图谱构建中的语义标准化问题.[方法/过程]文章在对国内外语义知识标准化相关内容研究的基础上,提出了领域知识图谱构建中应当遵循的五层标准体系模型,该模型融合了国内外领域相关的元数据与本体标准,揭示出知识图谱构建中以元数据标准为核心的重要思想,指导领域选择标准词汇的方法,使得领域所构建的知识图谱实现标准化,从而为知识的共享与互操作奠定基础.[结果/结论]在石油术语知识图谱构建过程以该标准体系模型作指导,实现了石油术语知识图谱标准构建,从而验证了标准体系模型的合理性与正确性,文中提出的观点和标准体系模型具有创新性.

教育是典型的知识密集型领域,因此,进行领域知识的共享研究对教育领域的发展非常重要.多年信息化的历史经验告诫我们:共享的前提是必须对领域资源实施标准化.尽管我国教育领域研发了系列教育元数据标准,但是从智慧教育的需求来看,这些标准普遍表现为缺乏语义、比较零散、体系不够完善等.随着近些年知识图谱的提出,为教育资源的知识整合发展提供了重要的技术手段.然而,目前知识图谱构建的主要焦点在构建技术方面,忽视了构建中的标准化问题.长此以往,领域将会构建出很多"知识孤岛".为解决这些问题,本文在对国内外教育领域内标准进行研究的基础上,以系统理论为依据提出并构建了一个集预定义词汇的W3C语义Web标准、顶层词汇标准和领域词汇标准于一体的三层架构模型——教育资源知识图谱标准词汇参考模型.基于该模型建构知识图谱,不仅为教育领域数据的标准化提供标准词汇参考,而且也为规范构建教育领域知识图谱提供了参考.