Mount Sinai finds deep learning algorithms inconsistent when applied to outside imaging data sets

The team used convolutional neural networks (CNNs) on its own chest X-rays as well as imaging from NIH and Indiana University and said they did not perform equally.