Genetic Programming as an Innovation Engine for Automated Machine Learning: The Tree-Based Pipeline Optimization Tool (TPOT)

One of the central challenges of machine learning is the selection of methods for feature selectionFeature selection, feature engineeringFeature engineering, and classification or regression algorithms for building an analytics pipeline. This is true for both novices and experts. Automated machine learning (AutoML)AutoML has emerged as a useful approach to generate machine learning pipelinesMachine learning pipelines without the need for manual construction and evaluation. We review here some challenges of building pipelines and present several of the first and most widely used AutoMLAutoML methods and open-source software. We present in detail the Tree-based Pipeline Optimization Tool (TPOT)TPOT that represents pipelines as expression trees and uses genetic programming (GP) for discovery and optimization. We present some of the extensions of TPOTTPOT and its application to real-world big data. We end with some thoughts about the future of AutoMLAutoML and evolutionary machine learning.