3 ms·
There was a thread on reddit the other day about what problems a NN could solve that an SVM couldn't. Apparently there are things such as structured SVM's and
by wodenokoto 10y ago
There was a thread on reddit the other day about what problems a NN could solve that an SVM couldn't.
Apparently there are things such as structured SVM's and SVM networks, and models where they attach an SVM instead of the output neurons.
It isn't dead, and people are coming up with some pretty cool stuff, but NN seems to be where it is at right now.