Events Calendar

CS Colloquium: Hyun Oh Song (Stanford) -Beyond supervised pattern recognition: Efficient learning with latent combinatorial structure

Speaker: Hyun Oh Song, Stanford

Talk Title: Beyond supervised pattern recognition: Efficient learning with latent combinatorial structure

Series: CS Colloquium

Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium

Supervised pattern recognition with 10^6 training data and 10^9 layered parameters has brought tremendous advances in artificial intelligence. However, there are two main limitations to this approach: 1) The knowledge learned in one area doesn't easily transfer to another and 2) supervising every single task is not only infeasible but also requires huge amounts of human labeled data which is costly and time consuming. In this talk, I will suggest a unifying framework which jointly reasons the prediction variable and the underlying latent combinatorial structure of the problem as a way to address such limitations. To demonstrate the practical benefits of the approach, we explore classification, localization, clustering, and retrieval tasks under settings that go beyond fully supervised pattern recognition.

Biography: Hyun Oh Song is a postdoc in SAIL in the computer science department at Stanford University. He received Ph.D. in Computer Science at UC Berkeley in late 2014 under the supervision of Prof. Trevor Darrell. He is a recipient of five year Ph.D. fellowship from Samsung Lee Kun Hee Scholarship Foundation. His research interest lies at the intersection between machine learning, computer vision, and optimization. He has an academic website at http://ai.stanford.edu/~hsong.

Host: CS Department

Webcast: https://bluejeans.com/501895444

Location: Henry Salvatori Computer Science Center (SAL) - 101

WebCast Link: https://bluejeans.com/501895444

Audiences: Everyone Is Invited

Contact: Assistant to CS chair