Data Science and Systems Engineering

Systems Conversation with S. Joe Qin, Chair Professor, Dean of the School of Data Science, and Director of Hong Kong Institute for Data Science, City University of Hong Kong

This Systems Conversation was conducted before S. Joe Qin gave a talk in our systems seminar series. Title of his talk: Dynamic Latent Predictive Analytics for Systems Data

Machine learning and modern statistics are major pillars of data science, which has enjoyed tremendous advances in both theoretical development and adoptions in the society. As data-driven technologies move into the domain of industrial and engineering systems, systems theory and engineering could be another pillar and a great application domain of data science. This talk addresses the challenge in modeling multi-dimensional time series data from engineered and natural systems, where dynamics are contained in a latent subspace with much reduced dimensions. The dynamic latent space is the dynamic kernel that facilitates predictive analytics. Industrial and synthetic datasets are used to illustrate this point of view to show unparalleled effectiveness over traditional methods. Related methods from subspace system identification, statistics, econometrics, and machine learning are mentioned to bring thoughts for future development of latent dynamic system analytics.