What is the difference between a hyperparameter search using random search versus grid search?

Question

Machine Learning — Hard

What is the difference between a hyperparameter search using random search versus grid search?

Accepted Answer

Grid search exhaustively evaluates all combinations in a predefined discrete grid; random search samples hyperparameter configurations randomly — random search is more efficient in high-dimensional spaces. This is because not all hyperparameters are equally important, and random samples cover the space more evenly per evaluation.