Session Overview
Solving high-dimensional optimization problems remains one of the key components in many applications, including design optimization, operations research, scientific exploration and so on. Traditionally these applications rely on heavy human experience to find good solutions and/or to tune the existing system for better performance. In this talk, I will cover our recent works in which deep neural networks, coupled with reinforcement learning and search methods, are used to learn heuristics of a complicated optimization problem, to achieve better performance than human experience. The application includes online job scheduling, neural architecture search, black-box optimization, and so on.
Overview
-
1
Learning to Optimize High-Dimensional Optimization Problems
-
Abstract & Bio
-
Learning to Optimize High-Dimensional Optimization Problems
-