Speaker: Dr. Sehun Chun Venue: B105 Date and Time:
About the Speaker
- (Current) Assistant Professor of Applied Mathematics, UIC, Yonsei University, Incheon - Residential Researcher, African Institute for Mathematical Sciences, Cape Town, South Africa - Research Associate, Imperial College London, London, UK - Research Assistant, Brown University, Providence, US - Ph.D. in Applied Mathematics, Brown University - M.S. in Applied Mathematics, Brown University & in Mathematics, Purdue University - BS in Mech. & Aero Engineering, Seoul National University - Research interest: Electrodynamics, Moving Frames, Curved surfaces, Electrophysiology, Cardiology, Neuroscience
Abstract
Recent advances in artificial intelligence (AI) have drawn unprecedented attention from every corner of our society. This presentation focuses on AI based on neural network algorithms. These algorithms derive from the study of a small region or function of the brain, particularly regarding vision. Their implications are almost limitless in the modern era but only a handful of applications survive, leaving other applications incomplete or inferior. Thus, we should pay attention to what is missing in AI algorithms compared to the real human brain before being overwhelmingly pessimistic about AI. Among the critical factors which will limit AI functionalities, we focus on the unique structure and distribution of neural fibers within a small volume of the brain, called the geometry of intelligence. The geometry of intelligence may explain the secret behind superior data processing efficiency and the mysterious entanglement with our physical world. The study of the geometry of intelligence can shed light on a possible breakthrough with respect to the current critical flaws of AI, or at least help us better understand the boundaries of AI for constructive and safe applications.