WASP Package: Bayesian optimization methods and their applications to real-world problems

Optimization problems are black-box optimization problems, where the objective function does not have an analytical expression, nor do we know its derivatives. Evaluation of the function is restricted to sampling and getting a possibly noisy response. If function evaluation is expensive, e.g. tuning hyperparameters of a deep neural network, it is important to minimize the number of samples drawn from the black box function. This is the domain where Bayesian optimization techniques are most useful because they attempt to find the global optimum in a minimum number of steps. Bayesian optimization incorporates prior belief about the black-box function and updates it with the samples drawn to get a posterior that better approximates the function. Our research is in Bayesian optimization methods and their applications to real-world problems such as hardware design, automated machine learning, computer vision, robotics, and database management.

The project is financed by Wallenberg AI, Autonomous Systems and Software Program (WASP)