[3Yin2] Interactive session 1

[3Yin2-14] Proposal of Food Information Knowledge Graph and AI Chef Machine for Digital Food Design.

〇Akio Kobayashi¹, Tetsuo katsuragi¹, Kengo Itoh¹, Motoko Inatomi¹, Keita Yamazaki², Takahiro Kawamura¹ (1.Research Center for Agricultural Information Technology, NARO, 2.University of Tsukuba)

Keywords:Knowledge graph, Ontology, AI chef machine, Health care

The Moonshot research and development program, "Development of innovative food solution for simultaneous food loss reduction and QoL improvement," aims to realize an AI chef machine that can automatically serve delicious and effective dishes to improve each individual's constitution and physical condition. To this goal, NARO constructs a knowledge graph that appropriately structures nutritional and functional food data, cooking recipe data, profiles, and so forth, and develops a system that outputs recipe data tailored to each person's preferences and physical condition to a 3D food printer. This knowledge graph links to FoodOn, a large-scale ontology of food, and can include knowledge that cannot be covered by the data analyzed and collected by the consortium, such as ingredients and cooking methods. As an inference method on knowledge graphs, we are investigating a method of probabilistically handling knowledge about food based on big data. In this paper, we describe the knowledge graph and its registered datasets and introduce experiments on the inference method on the knowledge graph.