Re:An Epidemic of AI Misinformation
by unishark on Saturday November 30, 2019 @08:09PM
(#21673135)
I think there are certainly issues with ML in radiology (i.e. lots of publications of low power studies or overfit models validated incorrectly), but I agree with you that Marcus was being unfair. It is principle a straightforward problem to solve with deep learning. The reason biggest reasons "no actual radiologists have been replaced" are probably just political/social. Hinton cannot be blamed for those hurdles.
[ Parent
| on HN ]
There are no comments on this comment yet.