Hosted by: Robert Harrison

Hypothesis driven modeling and its validation in the past took the reductionist approach where a scientist made a testable hypothesis. Such hypothesis driven analysis was the critical piece of scientific discovery. However, nowadays, due to the advancements of instrumentation, computational hardware and frameworks, we face a tsunami of data from the instruments or simulations and thus the hypothesis driven discovery has met serious challenges in scientific breakthrough. The information technologies addressed these issues using machine learning and data science approaches. In this talk, I will present the data driven modeling approach and show how machine learning can expedite the scientific discovery and share my vision for future data driven discovery.