Can AutoML outperform humans? An evaluation on popular OpenML datasets using AutoML Benchmark

In the last few years, Automated Machine Learning (AutoML) has gained much\nattention. With that said, the question arises whether AutoML can outperform\nresults achieved by human data scientists. This paper compares four AutoML\nframeworks on 12 different popular datasets from OpenML; six of them supervised\nclassification tasks and the other six supervised regression ones.\nAdditionally, we consider a real-life dataset from one of our recent projects.\nThe results show that the automated frameworks perform better or equal than the\nmachine learning community in 7 out of 12 OpenML tasks.\n