RBM architectures and applications

We will start with RBMs and their implementation, as well as their application in recommendation systems, and then move on to DBNs and utilizing them to classify images.

RBM and their implementation in TensorFlow

Let's begin with initializing the parameters for the RBM model. Recall that an RMB model consists of weights W associated with the connection between the visible layer and the hidden layer, bias a for the visible layer, and bias b for the hidden layer. An RBM object is configured by weights W, bias a and b, number of visible units and number of hidden units, number of Gibbs steps to take, besides regular neural network hyperparameters including batch size, learning rate, and...