AutoML on Azure Databricks automates machine learning pipelines. AutoML encompasses hyperparameter tuning and model search while providing data scientists with the flexibility and control they need.
Azure Databricks includes multiple tools for hyperparameter tuning. Hyperparameters are ML algorithm configurations that are traditionally tuned by hand. Common hyperparameters include regularization parameters, the number of epochs for deep learning training, and the depth of decision trees. Automated hyperparameter tuning uses grid search as well as more efficient algorithms such as Random Search and Bayesian-based algorithms. These automated methods are integrated seamlessly with MLflow, which makes tracking and managing your hyperparameters much easier.