| Abstract: |
Explosive progress made in the area of Artificial Intelligence and Machine Learning (AI/ML) has been adopted in many areas of research both in academia and industries. Backed up with ever-evolving computing processors, AI/ML techniques made "orders of magnitude" impact in the quality and speed of scientific applications. In the past decade, much of research focus has been importing AI/ML techniques from outside to simply engineer such individual applications. To reach (or to accelerate our reach) beyond where we are today, we need Scientific Machine Learning, an area of research in which we attempt to understand how AI/ML techniques works and improve further applying our knowledge from science domains. Introducing science knowledge, we can improve transparency and explainability of AI/ML applications, enable optimization of a workflow beyond individual applications, and expand the research impact beyond our domain science and into the area of AI/ML. In this talk, I will discuss recent progresses made in the Scientific Machine Learning, future directions, and the need of an ecosystem to accelerate this interdisciplinary research. |
