Speaker: Dr. Rak-Kyeong Seong of Samsung SDS

Abstract: Machine Learning has been a useful tool in recent years for solving computationally costly problems in physics and other natural sciences. However, often too much focus has been on what computation Machine Learning can do for us rather than on how AI can teach us about new physics and physical phenomena. By highlighting the strong connections that Machine Learning fundamentally has with Physics, the colloquium will outline the challenging problem of interpretability and explainability of AI and how we might tackle it in order to discover new physics. The colloquium will make connections to problems in industry, where explainable AI is quintessential for risk reduction, as well as to problems in a wide range of research areas such as molecular biology, medicine, electronic design automation or material science. At the heart of the colloquium will be our pioneering work in string theory since 2017 and how explainable AI is paving the way for new groundbreaking discoveries at the interface of theoretical physics and mathematics.

Bio: Dr. Rak-Kyeong Seong obtained his Ph.D. in Theoretical Physics at Imperial College London under the supervision of Professor Amihay Hanany. After postdoctoral fellowships in Korea and Sweden, he became a tenure-track assistant professor at Tsinghua University in Beijing where he stayed until 2019. Following his work on applying Machine Learning in string theory, he moved to Seoul to join the AI Advanced Research Lab at Samsung SDS, where he is as a Senior Researcher.