Research on Semantic Publishing Model Based on Knowledge Management Ecosystem

Abstract

With change of the scientific research paradigm and development of the semantic technology, the traditional publication is facing the need change of the users of scientific research, so it begins to transform into a knowledge service model, which traces back to the source of the scientific research innovation process, is based on the mass content resources of the authoritative systems, and provides diversified, three-dimensional, customized services and solutions for the researchers and organizations dealing with research project selection, literature retrieval and analysis, experiments, scholarly communication and exchange, accomplishment publication and scholarly evaluation. The knowledge management system is the foundation of knowledge service. The new technologies such as semantic web, artificial intelligence and knowledge mapping etc. also provide support and possibility for evolution of the knowledge management system. On basis of the above new technologies, a semantic publishing model is built in this study based on the knowledge management ecosystem. The knowledge management ecosystem using the machine readable RDF triple model as the metadata structure surpasses the restriction of the system or platform, makes it possible to understand each other between machines, and builds a cross-domain linking relationship which can be understood by the machine based on linked data and knowledge mapping so as to support and drive the transformation of science & technology publication into knowledge service, finally forming a machine-readable, data-driven, multidisciplinary collaboration, demand-oriented, resource deconstruction granularity, and gradually growing semantic publishing model. In the further study, a semantic publishing model was established on the basis of the knowledge management ecosystem in the field of life sciences and basic medicine. The resource of the model system included researchers, institutes, literature, dataset, patents, funds, honors and evaluation and so on, which were stored in RDF triples and published in form of the linked data, and combined with natural language processing (NLP) and neural network based on the deep learning technology so as to realize reasoning and linking of the new knowledge as well as complete knowledge growth with help of the machine intelligence. This model system could facilitate seamless integration between different databases and also promote cross-domain integration of the system with external system resources such as Pubmed, and construct the service modules such as knowledge acquisition, data mining, internalization, sharing, evaluation, and externalization and so on, in order to provide support guarantee for cooperation among multiple different agencies