chapter four

4 Automated generation of end-to-end ML solutions

 

This chapter covers

  • A brief introduction to AutoKeras
  • Automated classification and regression
  • Addressing multi-input and multi-output problems with AutoML

The previous chapters provided a basic introduction to machine learning, different kinds of ML models, and the workflow of handling ML problems. You’ve also already seen one of the most intuitive AutoML methods for hyperparameter tuning: using grid search to tune an ML pipeline with the help of the scikit-learn toolkit.

4.1 Preparing the AutoML toolkit: AutoKeras

4.2 Automated image classification

4.2.1 Attacking the problem with five lines of code

4.2.2 Dealing with different data formats

4.2.3 Configuring the tuning process

4.3 End-to-end AutoML solutions for four supervised learning problems

4.3.1 Text classification with the 20 newsgroups dataset

4.3.2 Structured data classification with the Titanic dataset

4.3.3 Structured data regression with the California housing dataset

4.3.4 Multi-label image classification

4.4 Addressing tasks with multiple inputs or outputs

4.4.1 Automated image classification with the AutoKeras IO API

4.4.2 Automated multi-input learning

4.4.3 Automated multi-output learning

4.5 Summary