The 3rd International Conference on Artificial Intelligence and Computer Engineering(ICAICE 2022)
Assoc. Prof. Xinying Chen

Assoc. Prof. Xinying Chen


Assoc. Prof.  Xinying Chen

School of Computer and Communication Engineering, Dalian Jiaotong University, China

Speech Title:


Multi-level semantic service structured network of semantic Internet of things


语义物联网是物联网内在矛盾的应对之策,它不是物联网和语义网的简单叠加,而是物联网的提升。语义互操作即在语义层面的信息交换和信息共享。语义物联网中的服务和其它服务并不是完全分裂的,网络中的服务加上语义标注后与物理实 体设备绑定映射,就具备了语义物联网服务的特点。 面对海量的、冗杂的网络数据,为在动态环境中,快速和准确地应对动态、复杂的服务需求,需要基于语义互操作对语义物联网服务的关键技术进行研究。目前,基于语义互操作的语义物联网服务的研究已经取得了一些进展,但仍然存在诸多亟待解决的关键问题,例如,如何有效解决语义关系的链接预测。我们针对这个热点问题,进行系统性的研究。为了解决语义物联网服务间的语义关系的链接预测和自动协作问题,提出了对应的解决方案,具体内容包括:1)首先,基于语义链接网络提出一种三层级语义服务结构化网络模型(SNM4SS)。SNM4SS 模型能够表达丰富的语义关系并进行推理,可以实现语义物联网服务的自动协作。2) 为了构建三层级语义服务结构化网络模型,基于 Markov 网、服务事件间的语义关系矩阵和边删除算法,分别提出基于条件互信息的语义服务事件网的构建算法(SSEN)和动态更新算法(SSEN_U)。3)由于语义服务事件网的动态构建需要解决服务事件链接推理问题,因此,基于 Markov 逻辑网 和随机游走策略,提出了一种服务事件链接推理算法(SELR)。SELR 推理算法能降低推理运算中的节点数,避免了利用闭 Markov 逻辑网建模所带来的巨大时空开销问题,可以更为有效解决面向多层级语义服务结构化网络中服务事件关系的链接预测问题。我们已通过一系列的实验验证了所提方法的有效性。

Semantic Web of Things (SWoT) is a solution to the internal contradictions of Internet of Things (IoT). It is not a simple superposition of IoT and Semantic Web, but an improvement on IoT. Semantic interoperability (semantic collaboration) is information exchange and information sharing at semantic level. SWoT services and other services are not completely split. Once the services in IoT are annotated with semantics and binded together with the mappings to their physical entity devices, which means, these services are possessed of the characteristics of SWoT services. Faced with massive and cumbersome network data, to respond quickly and accurately to dynamic and complex service requirements in IoT, some key correlated techniques on semantic interoperability based SWoT Services should necessarily be investigated. At present, some progress has been made on semantic interoperability based SWoT Services, but many key issues have still been unsolved, such as how to effectively solve the link prediction of semantic relationship. To solve the hot issue, we systematically focus on the following aspects, and the main research results are as follows. To solve the problems of link prediction and automatic collaboration on semantic relationship between SWoT services in distributed environment, the corresponding solutions are proposed. The specific contents include: 1) A three-level semantic service structured network model SNM4SS is proposed based on Semantic Link Network, which can express and reason rich semantic relations to collaborate SWoT services automatically. 2) In order to construct this model, based on Markov Network, semantic relationship matrix between service events and edge deletion algorithm, a Markov Network construction algorithm SSEN and a dynamic update algorithm SSEN_U based on conditional mutual information for semantic service events are proposed respectively. 3) Since the dynamic construction of semantic service event network needs to infer on service event link, a service event link inference algorithm SELR is proposed based on Markov Logic Networks and random walk strategy. This inference algorithm not only can reduce the node number in inference operation, which avoids huge space-time overhead caused by closed Markov Logic Networks modeling, but also can more effectively predict the links on service event relationship in the multi-level semantic service structured network. We verified the effectiveness of the proposed method through a series of experiments.



CHEN received the B.E. degree in computer science and technology and the M.E. degree in computer software and theory from Jilin University, China, in 2002 and 2005, respectively, and the Ph.D. degree in computer application technology from Dalian Maritime University, China. She is currently an Associate Professor with the School of Computer and Communication Engineering, Dalian Jiaotong University. Her current research fields include big data analysis, artificial intelligence, semantic Internet of things, computer vision, disease prediction, etc.