In this paper, we discuss artificial intelligence (AI) for science and one of its approach, data-driven science. Based on the tri-level of data-driven science proposed as basic theory, we show how to move ahead on AI for science. As a key issue to address in AI for Science, we introduce data-science framework to integrate extensive numerical data obtained by large-scale computation simulation, such as the “Kei(京)” computer, and by large-scale measuring system, e.g. synchrotron radiation and quantum beam. First, we propose Bayesian sensing, which is formulated base on the Bayesian inference, and Virtual Measurement Analysis (VMA) for analysis of the instrument data. Next, we introduce an extraction of effective model from electronic structure calculation for analysis of the simulated data. Finally, we discuss the integration of large-scale simulated and measurement data by data-driven approach through the effective model.
This paper presents a novel approach for real-time and proactive navigation in crowded environments such as event spaces and urban areas where many people are moving to their destinations simultaneously. The proposed approach is based on people flow predictions. In this paper, a new spatio-temporal statistical method based on machine learning technique is presented for the people flow prediction. The effectiveness of the proposed navigation approach by using computer simulation data using artificial people-flow data.