Java Deep Learning Essentials

Chapter 4. Dropout and Convolutional Neural Networks

In this chapter, we continue to look through the algorithms of deep learning. The pre-training that was taken into both DBN and SDA is indeed an innovative method, but deep learning also has other innovative methods. Among these methods, we'll go into the details of the particularly eminent algorithms, which are the following:

• The dropout learning algorithm
• Convolutional neural networks

Both algorithms are necessary to understand and master deep learning, so make sure you keep up.

Deep learning algorithms without pre-training

In the previous chapter, you learned that layer-wise training with pre-training was a breakthrough for DBN and SDA. The reason why these algorithms need pre-training is because ...