This chapter begins by teaching you how to create an end-to-end deep learning solution without selecting or tuning any deep learning algorithms. This can be done with as few as five lines of code, which is much simpler than the process introduced in chapter 3 for implementing a deep learning pipeline. Then you’ll learn how to perform classification and regression on image, text, and tabular data, as we did in the previous chapters, but with AutoML. We’ll also explore several more complex scenarios, including tasks with multiple types of inputs, such as both images and text, and tasks with multiple targets, such as a joint prediction of regression responses and classification labels.
Before starting to work on the real problems, let’s first explore our primary tool for AutoML, AutoKeras. AutoKeras is a Python library focused on the automated generation of deep learning solutions. To install AutoKeras, you can simply run pip install autokeras at your command line or !pip install autokeras in a Jupyter notebook. A more detailed discussion of the package installation is given in appendix A.