ANN/Deep Learning/CNN etc., are they not overhyped? Today under the sun they promise to deliver solution to every problem. But in practise they have limitations. Besides this they are not transparent in decisions. Convergence is a major issue. Continuous learning requires retraining, which is cumbersome. To me it looks they are overhyped. Other algorithms also perform equally well in decision making.
I think they are not over-hyped because these methods are getting results that were not possible before with classical programming.
Thanks Bryar Shareef, for your view point. Result in terms of accuracy may have improved for applications related to image classification. It does help in problems which fits in the auto-encoder mode for completely connected NN. But again it has issues with overfitting, error rates, convergence, and above all learning from new information (incremental learning). Transfer learning is not the true learning of concept being applied in recent works. I appreciate your response.
Indeed the learning process is not impartial via equally likely likelihood scale of mathematical statistics. Its pleaders claims convolution can do the only good job, which is only because of its simplicity, but what about saddle point approximations and Laplace approximations of high dimensional integrals which produces the ever best results ever????
CNN and ANN , some best classifiers in terms of accuracy, time, and the amount of data they can handle.
But you can say, there is no perfect classifier because there is no perfect training data.
- Badges
- Science topic