Papers: Metadata as Linked Data and Knowledge Graph

In the World Wide Web, a very large number of online cultural heritage (CH) resources is made available through digital museums websites. They display collections images and share collection metadata, which makes opportunities for aggregating. 'China Art in the Museums Overseas' platform aggregated China digital collections from overseas institutions. This paper introduces the metadata aggregation workflow of this project, especially a unified data model used to solve the metadata standards heterogeneity. Besides, this platform is not simply a database, but also provides search, images semantic annotation, knowledge graph serves, linked open data interface and so on. Our work is an effort to resolve multi-source data aggregation and it is valuable. It improves the availability and discoverability of digital collections, and spread Chinese culture to the public more widely.

In the process of rapid urbanization in China, the protection and utilization of great sites are facing unprecedented pressure. The effective knowledge organization of great sites is a prerequisite for their protection and utilization. Ontology provides realization paths for organizing knowledge of great sites. In this paper, firstly, CIDOC- CRM and Time ontology are reused to build the top-level ontology with the results of user interviews. Secondly, top-down concept extraction and bottom-up concept expansion are adopted to gain the knowledge concepts and instances of great sites. Finally, data properties and object properties are defined. The designed ontology can provide the knowledge modeling and representation of great sites, laying the foundation for knowledge sharing.

Application Profiles are a collection of vocabularies mixed and matched from different namespaces and customized for the local application. Application Profiles act as a constrainer as well as an explainer of the metadata for every dataset. Linking data and generating knowledge graphs are the general challenges that information processing communities are trying to address constructively. Application Profiles can act as a means of linking data and generating knowledge graphs. The authors propose application profile driven questionnaire creation for data linking in the perspective of crowdsourced data acquisition; especially where there are challenges in adapting a single vocabulary or a limited number of adaptable domain-specific vocabularies. This paper presents a proof of concept study based on the proposed approach by adapting existing standards and tools, with the notion that similar methods are applicable in related use-cases.