3 ms·
Neuroimaging has always had an issue with interpretability and susceptibility to statistical artifacts. I fear that moving to another black box is only going to
by spupe 4y ago
Neuroimaging has always had an issue with interpretability and susceptibility to statistical artifacts. I fear that moving to another black box is only going to complicate matters.
- exdsq 4y agoFWIW my wife’s a researcher at Stanford applying ML/DL to MRI to speed up acquisition times and working closely with clinicians (I think they’re actually rolling it out in the hospitals) and the results have all been positive.