@INM7_ISN 2 years ago
Do we need to re-think #MachineLearning for (brain) medicine? Current proof-of-concept studies on retrospective, repository data may not yield a relevant practical impact @den_hed & I argue for more "doctors at the black box" for more user perspective https://www.nature.com/articles/s41380-020-00931-z https://t.co/6NJrne9tuu
@INM7_ISN 4 years ago
Eysenk's astonishing work on "cancer personality". If results are too good to be true, there is usually a good reason for this Highly relevant to current work on #MachineLearning in neuroimaging and Medicine. Real life is too complex for >>90% accuracy https://cosmosmagazine.com/society/is-this-one-of-the-worst-scientific-scandals-of-all-time