6 AutoML with a fully customized search space

 

This chapter covers

  • Customizing the entire AutoML search space without connecting AutoML blocks
  • Tuning autoencoder models for unsupervised learning tasks
  • Tuning shallow models with preprocessing pipelines
  • Controlling the AutoML process by customizing tuners
  • Joint tuning and selection among deep learning and shallow models
  • Hyperparameter tuning beyond Keras and scikit-learn models

This chapter introduces customization of the entire AutoML search space in a layerwise fashion without wiring up AutoML blocks, giving you more flexibility in designing the search space for tuning unsupervised learning models and optimization algorithms. We introduce how to tune a shallow model with its preprocessing pipeline, including feature engineering steps. You will also learn how to control the model-training and evaluation process to conduct the joint tuning and selection of deep learning models and shallow models. This allows you to tune models with different training and evaluation procedures implemented with different ML libraries.

6.1 Customizing the search space in a layerwise fashion

6.1.1 Tuning an MLP for regression with KerasTuner

6.1.2 Tuning an autoencoder model for unsupervised learning

6.2 Tuning the autoencoder model

6.3 Tuning shallow models with different search methods

6.3.1 Selecting and tuning shallow models

6.3.2 Tuning a shallow model pipeline

6.3.3 Trying out different search methods

6.3.4 Automated feature engineering

6.4 Controlling the AutoML process by customizing tuners

6.4.1 Creating a tuner for tuning scikit-learn models

6.4.2 Creating a tuner for tuning Keras models

6.4.3 Jointly tuning and selection among deep learning and shallow models

6.4.4 Hyperparameter tuning beyond Keras and scikit-learn models