Azure ML Studio AutoML Hyperparameter Optimization & Algorithm Selection

Question

Which method (Grid Search, Bayesian Search, Random Search, ...) is used in Azure ML Studio AutoML per default to optimize model hyperparameters in order to increase model accuracy? In the SDK you can choose which method you want to use (according to https://learn.microsoft.com/en-us/azure/machine-learning/how-to-tune-hyperparameters), but I could not find any information about the Studio execept another asked question, but the provided answer was not clear to me (https://learn.microsoft.com/en-us/answers/questions/462352/hyperparamter-optimization-in-azure-automl.html).

And how exactly does Azure ML Studio AutoML search for possible algorithms/ how does the tool choose, which algorithm to use next?

Answer 1

Hi, according to this document, Azure AutoML uses Bayesian optimization to optimize hyperparameters. How Azure AutoML works - During training, Azure Machine Learning creates a number of pipelines in parallel that try different algorithms and parameters for you. The service iterates through ML algorithms paired with feature selections, where each iteration produces a model with a training score. The higher the score, the better the model is considered to "fit" your data. It will stop once it hits the exit criteria defined in the experiment. Hope this helps!