207. Improving the robustness of zeroth-order optimization solvers
Invited abstract in session MC-1: Strategies to Improve Zeroth-Order Optimization Methods, stream Zeroth and first-order optimization methods.
Monday, 14:00-16:00Room: B100/1001
Authors (first author is the speaker)
| 1. | Stefan M. Wild
|
| Applied Mathematics and Computational Research Division, Lawrence Berkeley National Laboratory |
Abstract
Zeroth-order optimization solvers are often deployed in settings where little information regarding a problem's conditioning or noise level is known. An ideal solver will perform well in a variety of challenging settings. We report on our experience developing adaptive algorithms, which leverage information learned online to adapt critical algorithmic features. We illustrate our approach in trust-region-based reduced-space methods and show how trained policies can even be deployed effectively in nonstationary cases, where the noise seen changes over the decision space. This is joint work with Pengcheng Xie.
Keywords
- Derivative-free optimization
- AI based optimization methods
- Optimization under uncertainty
Status: accepted
Back to the list of papers