Mickaël Binois, a postdoctoral appointee in Argonne’s Mathematics and Computer Science Division, will give a presentation at the Argonne Machine Learning/High Performance Computing Journal Club on January 16, 2019. 

In his presentation, titled ​“Directions for High-Dimensional Bayesian Optimization,” Binois will discuss several options that researchers can use to cope with tens to thousands of variables in Bayesian optimization. Particular emphasis will be given to the method of active subspaces,  which has been used to exploit the low-dimensional structure in certain problems, including those from climate and aerospace.

The club meets monthly at Argonne to highlight the applications and methods using high-performance computing for machine learning.