Future of automated machine learning

It has been over 5 years since automated ML tools started appearing in the data science community. There are now several open source tools (for example, TPOT and Auto-WEKA) as well as proprietary tools (for example, Kortical and H2O) on the market. All major cloud providers now have some AutoML offering. Interest in these tools has been rising. It is safe to say that interest in these tools will keep rising and more offerings should be expected to come onto the market. We expect the tools to keep expanding the scope of tasks that will be automated to cover more aspects of the model development process. We can already see tools such as DataRobot offering more functions through internal development as well as through acquisitions such as Paxata. While it remains true that these tools do not support all use cases or types of modeling, they do cover a large number of use cases in a typical organization. More capabilities and algorithms are being added...