Graph data management and analysis

Graph data can model the relationships between entities and is used in various familiar applications. For instance, by structuring the relationships between objects as knowledge graphs, they are utilized in web searches and product recommendation systems. Additionally, molecular data can represent atoms as entities and the connections between atoms as relationships, which can be used for searching for molecules with the same data structure. Social networking services and road networks are also closely integrated into our lives, leading to the growth and diversification of graph data.

In terms of graph data, I am advancing research and development from four perspectives: management, search, discovery, and prediction. Specifically, we are developing (1) database management techniques enabling efficient management and rapid searches, (2) data mining techniques to discover new insights, and (3) deep learning techniques for accurate predictions.

For example, I have developed techniques to speed up queries in graph databases and to extract the strengths or distinctive relationships in graph data, allowing us to evaluate if there are exceptional relationships or discriminative biases within the graph data. I am also working on deep learning algorithms applicable to large-scale graph data and algorithms that can automatically select graph nerual network architectures.

Mobile and spatio-temporal data analysis and management

With the advancement of the Internet of Things and location-based services, many data now come with both time and location information. For instance, data on human flow, traffic information, restaurant details, and sensor data are typical examples. Nowadays, many people are using services that handle spatio-temporal data, often without even realizing it.

For mobile and spatiotemporal data, I am advancing research and development from four perspectives: management, search, discovery, and prediction. For example, we're developing technologies for the acceleration of spatial queries (like searching for data within a certain range or finding the data closest to a certain point), optimization of distributed parallel processing, data mining to detect relationships between data, and prediction techniques using deep learning for sensor data values.

Furthermore, in mobile and spatio-temporal data analysis, we can apply graph data analysis techniques to capture correlations based on road networks or distances.

Interdisciplinary applications

Data science and artificial intelligence technologies are spreading widely throughout society. Even in fields that had not previously utilized information processing technology, they use such tecnologies, leading to collaborations with researchers from many diverse disciplines. For example, in chemistry and materials science, we try to discover new useful substances. In medicine, we analyze and predict patients with urinary stones. In philosophy, we explore the issues of AI and its societal applications, and in urban engineering, we analyze traffic volume and energy demand. By applying core information processing technologies to a wide range of fields, we acquire new knowledge outside of informatics and understand the new technologies demanded by society, which aids in conceiving the technologies we should research and develop.