After we finished training, as we can see from the following screenshot, we get a result of over 99%, so that is significantly better than what we got with softmax or our SVM:
Deep learning is probably the most powerful machine learning technique, due to its ability to learn very complex pattern recognition. It's just dominating everything else for advanced computer vision, speech processing, and more—stuff that conventional machine learning techniques haven't been all that successful at. However, that doesn't necessarily mean that you want to use deep learning for everything. Deep learning generally acquires a large number of examples—many thousands, if not millions sometimes—and it can also be very computationally expensive. So, it's not always the best solution, although it is very powerful, as we have seen right...