To analysis for food statistical surveys by the National Institute of Biomedical Innovation, Health and Nutrition (NIBIOHN) and other related institutes, the food ontology FGNHNS was constructed from the food groups of the National Survey of Nutrition and Dietetics, and is now available on BioPortal. As further extensions, we are adding Wikidata information, linking with FoodOn, integrating with the Standard Tables of Food Composition of Japan, linking with the Crop Vocabulary (CVO), and linking with food allergy information. We will report the progress at the workshop.
Wikidata, one of the largest knowledge graphs curated by the worldwide community, has been growing enormously over the years, leading to many concerns about data quality issues. Much factual knowledge of Wikidata lacks references information or insufficient information to support other fact- checking applications. In this paper, we first analyze the provenance status of researchers of Wikidata as a topical Wikidata domain knowledge. Second, we propose a new evaluation method to judge the provenance information of the target domain. Using our new evaluation methods, we found many references in the researcher domain inaccessible or difficult to access.
Causal relation knowledge is necessary to develop a facilitator agent that can understand discussion points and participants' opinions. However, it is not enough to be included in the Knowledge Graph. In this study, we attempted to extend the training data using Wikidata's casual relation knowledge as a method for extracting causes. To compare whether the proposed extraction method is more accurate than previous methods, we compared the accuracy of the output causes by inputting sentences. In addition, a calculation method was examined to determine if the extracted causes could be considered a general causal relationship. As a result, the accuracy of the extraction is improved over conventional methods, and a threshold value can be determined to consider it as a general causal relation. Future work includes the development of a facilitator agent to support discussions using the methods in this paper.
In recent years, the digital twin has emerged in the fields of building automation systems (BAS) and factory automation (FA). It is expected that information in real space can be reproduced in digital space to optimize control and management. On the other hand, data management of physical assets in each control and management paradigm has problems such as heterogeneity of data models and silos. To solve these problems, data modeling utilizing semantics has been attracting attention. In this study, we proposed and evaluated an ontology for representing digital twins for indoor spaces.
This paper describes the configuration of a regional information support system that links the collected data with DBpedia, Wikidata, etc., as information has become big data and open data with the recent development of Internet Web technologies. The databases of current Internet sites are becoming sophisticated and semantic, and are being constructed as knowledge bases, which are used for searches and the like, and can also provide a variety of related information. This system collects information on various facilities, offices, shops, etc. in the Sendagaya area, Shibuya-ku, Tokyo, organizes them into three groups of subject, predicate, and object, and converts them into RDF (Resource Description Framework) data. In addition, by linking with information posted on DBpedia, etc., we support the provision of information such as tourism that users want.
This paper describes methods for extracting and summarizing information that changes with time in a semantic coherent manner so that people can easily understand what has changed. Specifically, using CTD (Comparative Toxicogenomics Database) as an example, we describe the extraction of differences in CSV and some KG representation, difference extraction, and summarization with version information. We also refer to the application to laws and regulations.