Spatio-temporal Causal Modeling for Reliable Decision-Making

Machine learning and data mining, a field of artificial intelligence research, have dramatically improved performance in the tasks of knowledge discovery and prediction by adopting big data and deep learning methods. The spatio-temporal data analysis, which is a technology for processing big data acquired by IoT/5G, has become a major technological topic in artificial intelligence for decision-making in world society.

If AI provides accurate predictions, working with AI is expected to contribute to decision-making and consensus building process. On the other hand, if the analysis results are inaccurate, there is a risk of misdirected decision-making, and problems are expected to occur in society in the future. In this project, we are focusing on spatio-temporal bias of data, which is the cause of inaccurate AI analysis results, and attempt to develop novel spatio-temporal causal inference models to realize robust and reliable analysis against this problem.