Knowing The What But Not The Where in Bayesian Optimization
This work addresses a specific, incremental improvement in Bayesian optimization for applications where optimum outputs are pre-known, such as hyperparameter tuning in machine learning.
The paper tackles the problem of Bayesian optimization when the optimum output value is known but the corresponding input is unknown, proposing two new acquisition functions that leverage this prior knowledge. The methods demonstrate quantitatively better performance than standard Bayesian optimization in tuning deep reinforcement learning on CartPole and XGBoost on the Skin Segmentation dataset.
Bayesian optimization has demonstrated impressive success in finding the optimum input x* and output f* = f(x*) = max f(x) of a black-box function f. In some applications, however, the optimum output f* is known in advance and the goal is to find the corresponding optimum input x*. In this paper, we consider a new setting in BO in which the knowledge of the optimum output f* is available. Our goal is to exploit the knowledge about f* to search for the input x* efficiently. To achieve this goal, we first transform the Gaussian process surrogate using the information about the optimum output. Then, we propose two acquisition functions, called confidence bound minimization and expected regret minimization. We show that our approaches work intuitively and give quantitatively better performance against standard BO methods. We demonstrate real applications in tuning a deep reinforcement learning algorithm on the CartPole problem and XGBoost on Skin Segmentation dataset in which the optimum values are publicly available.