Statisticians are cautioning fellow scientists not to make assumptions about the accuracy, uncertainty or reproducibility of scientific discoveries made with today’s machine learning models.

Original source: https://www.sciencedaily.com/releases/2019/02/190215110303.htm