3 ms·
The problem is that the numerous easy cases do not result in a useful network ... any 1st-year intern will get the easy results already. The hard cases, which
by Ensorceled 8y ago
The problem is that the numerous easy cases do not result in a useful network ... any 1st-year intern will get the easy results already.
The hard cases, which would be useful to a doctor, occur very rarely. My father's unusual reaction to a post-bypass drug regimen was something like the 3rd time that happened in Canada. How do you "train" that into a neural network?
- draugadrotten 8y agoSounds like anomaly detection would be quite useful for monitoring the expected reactions and comparing with actual outcomes. It may not be the answer to "what is the right drug?" but it may be enough to say "what he got ain't right for him"
- marmaduke 8y agoI work on an ML project on epilepsy, and this happened all the time. Additionally what the clinicians saw as easy/boring wasn’t always te case for the model.