Updated on 2023-01-09 GMT+08:00

Bayesian Optimization (SMAC)

In Bayesian optimization, it is assumed that there exists a functional relationship between hyperparameters and the objective function. Based on the evaluation values of the searched hyperparameters, the mean and variance of the objective function values at other search points are estimated through Gaussian process regression. The mean and variance are then used to construct the acquisition function. The next search point is the maximum value of the acquisition function. Compared with grid search, Bayesian optimization uses the previous evaluation results to reduce the number of iterations and shorten the search time. The disadvantage is that it is difficult to find the global optimal solution.

Table 1 Bayesian optimization parameters

Parameter

Description

Recommended Value

num_samples

Number of times to sample from the hyperparameter space

The value is an integer ranging from 10 to 20. The larger the value, the longer the search time and the better the effect.

kind

Acquisition function type

This value is string type. The default value is ucb and other value options are ei and poi. Do not change the default one.

kappa

Adjustment parameter of acquisition function type ucb, which is the upper confidence boundary

This value is a float. You are advised not to change it.

xi

Adjustment parameter of acquisition function types poi and ei

This value is a float. You are advised not to change it.