'Geometry and Latent Representations in Machine Learning' 세미나 안내
세미나 안내드립니다. 관심있는 분들은 참여바랍니다.
· Title: Geometry and Latent Representations in Machine Learning
· Speaker: Prof. Daniel D. Lee (Cornell Tech)
· Date & Time: June 7 (Wed.), 16:00PM ~ 17:30PM (KST)
· Venue: Rm.1503 & Zoom Live Streaming
(https://us06web.zoom.us/j/
(Meeting ID: 876 3905 7456 / Password: 682464)
· Abstract:
The advent of deep neural networks has brought significant advancements in the development and deployment of novel AI technologies. Recent large-scale neural network architectures have shown significantly better performance for object classification, segmentation, scene understanding and multimodal representations. How can we understand how the representations of sensor input signals are transformed by deep neural networks? I will show how statistical insights can be gained by analyzing the high-dimensional geometrical structure of these representations as they are reformatted in neural network hierarchies.
· Bio:
Prof. Daniel D. Lee is Professor at Cornell Tech. Until recently, He was the Executive Vice President for Samsung Research and head of the Samsung AI Centers in seven locations worldwide. He received his B.A. summa cum laude in Physics from Harvard University and his Ph.D. in Condensed Matter Physics from the Massachusetts Institute of Technology. He was ranked in the PUTNAM top 25, which signifies an individual with exceptional mathematical problem-solving abilities. He served as the director of GRASP robotics lab at the University of Pennsylvania, which features a faculty of more than twenty world-leading robotics-AI-engineering professors. He won the robocup world-champion for many years. In Urban challenge with autonomous cars in 2007, he led one of the six teams who finished the race, among which his team was the only one to make it with small budget (Track B). He is a Fellow of the IEEE and AAAI and has received the National Science Foundation CAREER award and the Lindback award for distinguished teaching. In 2013, he received the NIPS (currently NeurIPS) classic paper award out of the nominated influential papers in the NIPS history by then. NeurIPS is the most prestigious conference in machine learning and artificial intelligence. He was the Program Co-chair and General Co-chair of NeurIPS conference in 2015 and 2016, respectively, and he is currently the board member of NeurIPS. His Nature paper in 1999 with title "Learning the parts of objects by non-negative matrix factorization" and the Science paper in 2000 with title "The manifold ways of perception" have pioneered new research fields in machine learning and theoretical neuroscience.