Chapter 6: Automated Machine Learning in Amazon SageMaker

Automated machine learning (AutoML) is the process of automating different aspects of the machine learning pipeline to help build and deploy high-quality models in a short period of time. This works by automating different phases of the machine learning process, such as feature engineering, architecture search, and hyperparameter optimization. Initially, tools for AutoML focused more on automating the time-consuming hyperparameter optimization tasks by looking for an optimal set of hyperparameters for a model. These past couple of years, however, AutoML has expanded to include the automation of other parts of the pipeline, including data cleaning, feature selection, model selection, and more:

Figure 6.1 – Different phases and types of tasks that are automated through AutoML

The preceding image shows a simplified diagram showing the different phases and types of tasks that can be automated with...