A Comparative Study Of An On Premise Automl Solution For Medical Image Classification

This article discusses the use of automated machine learning (AutoML) in medical applications, specifically in the field of medical imaging. The authors benchmark an open-source AutoML framework, AutoKeras, against bespoke deep learning architectures on five public medical datasets and find that AutoKeras can outperform the bespoke models in general. The article also explores the impact of common parameter choices on performance.