Manuals have traditionally been widely used in organizations as a means of educating and training new employees and sharing knowledge among staff. However, as the expression "manual man" implies, the negative effects of adhering so closely to written actions and rules that one is unable to respond appropriately when confronted with something that deviates from them have also been pointed out. We propose a knowledge-data structuring method that structures knowledge in a goal-oriented manner and combines it with data such as accident reports and best practices. This method enables us to understand the purpose and rationale of each action and to develop the ability to deal with new situations. However, the current construction method does not make use of the expertise of conventional experts in building high-quality manuals. Therefore, in this paper, a manual construction expert designs a care manual for self-reliance support for elderly based on structured knowledge by a knowledge engineering expert, and the expert's know-how is expressed by evaluating the quality of the manual at the time of design by users. The results of this study reveal the advantages and challenges of using structured knowledge as a knowledge source in manual construction, as well as changes in knowledge construction.

In recent years, researchers from the fields of computer vision, language, graphics, and robotics have tackled Embodied AI research. Embodied AI can learn through interaction with the real world and 3D environments and can perform various tasks in 3D environments using virtual robots. However, many of these are one-way tasks in which the interaction is interrupted only by answering questions or requests to the user. In this study, we aim to develop a task-oriented interactive system using 3D household ontology and commonsense reasoning, in which a virtual robot can reason about the location of a guide while interacting with the user and guide the user around the house.

CSQA is a question answering benchmark using the knowledge graph, that created using the large-scale knowledge graph database Wikidata. CSQA targets interactive question answering, and there are 10 question types, some of which require inference such as comparison and set operation. In the previous study, multi-task semantic parsing model that converts utterances to logical forms that expresses operations on a database can answer high accuracy. The conversion to a logical form is performed by defining the grammar for deriving the logical form and predicting the order in which the grammar is applied from the utterance. In order to learn a model for the prediction, it is necessary to search for a logical form that can answer the question from logical forms that can be generated by applying the grammar. However, since the search space is enormous, depending on the search method, problems such as a decrease in the success rate of the search and incorrect logical forms occur, which may adversely affect the learning of the model. In this research, we propose a method for searching logical forms with a high success rate of the search in a short time by using patterns of logical forms.

In both engineering and philosophical ontologies, there is no doubt about the importance of the two axes of Independent-Dependent (I-D axis) and Continuant-Occurrent (C-O axis) in their upper categories. However, the way these two axes are handled is not consistent among existing upper ontologies. In this study, we first showed that upper ontologies can be classified into three types based on the priority of the treatment of the two axes: C-O axis priority type, I-D axis priority type, and both axes equivalent type. Next, we clarified that the C-O axis priority type does not distinguish between Independent and Dependent of Occurrent, while the I-D axis priority type does not distinguish between Independent and Dependent of Occurrent.

Wikipedia is one of the largest knowledge resources available via the Internet. Several methods have been proposed for extracting knowledge/information from Wikipedia, but methods based on extracting knowledge from the Wikipedia category structure can not utilize whole part of its structure because of the complexity of the relationships between the various categories. In this paper, we briefly review our previous researches on analyzing Wikipedia category and introduce a framework called "Wikipedia Category Ontology" (WCO) that aims to act as a basis for interpreting the Wikipedia category structure. It is based on a classification of category types and relationship types and available online in the form of Linked Open Data at http://wcontology.org/. We also demonstrate the system by using Wikipedia category analysis scenarios.

Ontology development for a specific domain is a task to take a large cost. To reduce the costs, ontology development methods are needed. In this study, we investigate a semi-automatically method to extract concepts for domain ontologies from Wikidata, a large-scale public Linked Open Data (LOD) to develop a preliminary ontology. Our previous work extracted domain concepts with their conceptual hierarchies from Wikidata. In this study, we use these extracted concepts as input to extract conceptual relationships among domain concepts from Wikidata This paper discusses a method to extract relationships as triples related to domain concepts, and examines how extracted relationships could be used as conceptual relationships for the domain ontology.

In this paper, we describe a technique for enriching the ontology of background knowledge by connecting to wikidata in order to improve the accuracy of link prediction in knowledge graphs where entities are not typed in detail. We proposed an algorithm to remove entities that are irrelevant to the type of prediction from the training as negative examples. We confirmed that the proposed algorithm improves the accuracy compared to the case without connecting to the existing wikidata.