Reading questions

If you are unsure about any of these concepts, post a question on the discussion forum!

  • What is the energy function of an RBM?
  • What are the visible and hidden layers of an RBM?
  • What is the probability distribution of an RBM?
  • What are the conditional distributions of each layer given the other?
  • What loss do we try to optimize to train an RBM?
  • What is the Contrastive Divergence (CD) and how does it approximate the gradients used for training?
  • What is a Gaussian-Bernoulli RBM?
  • What is the free-energy of an RBM?
  • What is a Boltzmann machine (i.e. non-restricted)?
  • How de we perform Gibbs sampling in an RBM?
  • What is a negative sample?
  • What is pesistent CD?