XAIENCE 2019

Session

Invited talks - Math & Stats 1

7 Nov 2019, 09:20

국제회의실 (International Conference Hall) (Seoul National University, College of Natural Sciences, Building 25-1)

Conveners

Invited talks - Math & Stats 1

• Yongdai Kim (Seoul National University)

Contribution list Timetable

13. Theoretical Advantages of Deep Neural Networks

Prof. Yongdai Kim (Seoul National University)

07/11/2019, 09:20

From statistical points of view, deep neural networks (DNN) are nothing but a (generalized) regression model, but DNNs have solved many problems no other methods have not succeeded in the past. In this talk, I will explain theoretical advantages of DNNs compared to other nonparametric regression models.

12. Projection onto Minkowski Sums with Application to Constrained Learning

Prof. Joong-Ho Won (Seoul National University)

07/11/2019, 09:40

We introduce block descent algorithms for projecting onto Minkowski sums of sets. Projection onto such sets is a crucial step in many statistical learning problems, and may regularize complexity of solutions to an optimization problem or arise in dual formulations of penalty methods. We show that projecting onto the Minkowski sum admits simple, efficient algorithms when complications such as...