2447. Bayesian optimization for mixed feature spaces using tree kernels and graph kernels
Invited abstract in session WD-5: Semi-plenary talk Misener, stream PC Stream.
Wednesday, 15:15-16:00Room: H7
Authors (first author is the speaker)
| 1. | Ruth Misener
|
| Department of Computing, Imperial College London |
Abstract
We investigate Bayesian optimization for mixed-feature spaces using both tree kernels and graph kernels for Gaussian processes. With respect to trees kernels, our Bayesian Additive Regression Trees Kernel (BARK) uses tree agreement to define a posterior over sum-of-tree functions. With respect to graph kernels, our acquisition function with shortest paths encoded allows us to optimize over graphs, for instance to find the best graph structure and/or node features. We formulate both acquisition functions using mixed-integer optimization and show applications to a variety of challenges in molecular design, engineering and machine learning.
The tree kernel work is joint with Toby Boyne, Alexander Thebelt, Jose Folch, Calvin Tsay, Robert Lee, Nathan Sudermann-Merx, David Walz, and Behrang Shafei.
The graph kernel work is joint with Yilin Xie, Shiqiang Zhang, Jixiang Qing, and Calvin Tsay.
Keywords
- Stochastic Models
- Mixed-Integer Programming
Status: accepted
Back to the list of papers