Sander Lamballais presented a controversial paper showing no benefit of (complex) machine learning over logistic regression when predicting clinical outcomes.

Thanks for the interesting discussion!