How to Use
First, initialize an RBM with the desired number of visible and hidden units.
rbm = RBM(num_visible = 6, num_hidden = 2)
Next, train the machine:
training_data = np.array([[1,1,1,0,0,0],[1,0,1,0,0,0],[1,1,1,0,0,0],[0,0,1,1,1,0], [0,0,1,1,0,0],[0,0,1,1,1,0]]) # A 6x6 matrix where each row is a training example and each column is a visible unit.
r.train(training_data, max_epochs = 5000) # Don't run the training for more than 5000 epochs.
Finally, run wild!
# Given a new set of visible units, we can see what hidden units are activated.
visible_data = np.array([[0,0,0,1,1,0]]) # A matrix with a single row that contains the states of the visible units. (We can also include more rows.)
r.run_visible(visible_data) # See what hidden units are activated.
# Given a set of hidden units, we can see what visible units are activated.
hidden_data = np.array([[1,0]]) # A matrix with a single row that contains the states of the hidden units. (We can also include more rows.)
r.run_hidden(hidden_data) # See what visible units are activated.
# We can let the network run freely (aka, daydream).
r.daydream(100) # Daydream for 100 steps on a single initialization.
Introduction
Suppose you ask a bunch of users to rate a set of movies on a 0-100 scale. In classical factor analysis, you could then try to explain each movie and user in terms of a set of latent factors. For example, movies like Star Wars and Lord of the Rings might have strong associations with a latent science fiction and fantasy factor, and users who like Wall-E and Toy Story might have strong associations with a latent Pixar factor.
Restricted Boltzmann Machines essentially perform a binary version of factor analysis. (This is one way of thinking about RBMs; there are, of course, others, and lots of different ways to use RBMs, but I'll adopt this approach for this post.) Instead of users rating a set of movies on a continuous scale, they simply tell you whether they like a movie or not, and the RBM will try to discover latent factors that can explain the activation of these movie choices.
More technically, a Restricted Boltzmann Machine is a stochastic neural network ( meaning we have neuron-like units whose binary activations depend on the neighbors they're connected to; meaning these activations have a probabilistic element) consisting of:
