Due to the growing expansion of scientific literature and knowledge, access to scholarly contents is becoming more challenging. This paper presents a research profiling framework driven by scientific knowledge graphs (SKGs), which aims at achieving the deep fusion and thorough disclosure of scientific resources and domain knowledge, as well as satisfying the needs of researchers for knowledge overview and acquisition in a reasonable period of time using fine-grained and multi-dimensional profiling scenarios. Further, we conduct the experiment of the SKG construction task on a specific domain by combining scientific literature from Web Of Science and an open knowledge base, and develop a prototype system for user's single entity search to validate the approach, with profile presentations of overall graph view, literature profile, hotness topics list, high-impact experts, and providing a service of profile viewing and download.

This paper is to gain an overall view of the current research status of semantic retrieval systems based on knowledge graph, and construct such a system of rice field. The authors selected several well-known semantic retrieval models of various industries, such as Google, Bing, Elsevier, Baidu ZhiXin, etc., and comprehensively analyzed them from the aspects of entity type, field, entity source, retrieval mode, and information presentation, and then proposed a semantic retrieval system on basis of ontology, knowledge graph and nature language processing with multi-source of rice. The survey result shows that semantic retrieval is stepping toward the multi-field and instantiation stably, and the future research is concerned with semantization depth and multi-field knowledge graph. The proposed system can provide services of entity and knowledge retrieval, visual navigation and analysis, which meet users' retrieval needs in semantic to a certain extent.

Current solutions that link research data with scientific literature are primarily constrained in the dimension and depth of relations, which is a necessity for the sharing and reuse of scholarly resources and accelerating scientific progress. This paper presents a data-literature linking framework that combines data enrichment techniques of entity extraction, discipline classification & subject indexing based on semantic knowledge bases, and further implement multi-scenario data-literature link discovery with data from persistent identifier provisions - DataCite and CrossRef, including research data retrieval and faceted browsing, data-literature and data-data bidirectional discovery, entity-based link browsing and navigation.