This network, also known as Harmonium, was proposed in Information Processing in Dynamical Systems: Foundations of Harmony Theory, Parallel Distributed Processing, Vol 1, The MIT Press, 1986) as a probabilistic generative model. In other words, the goal of an RBM is learning an unknown distribution (that is, a data-generating process), in order to generate all possible samples. The generic structure is shown in the following diagram:
Structure of a generic Restricted Boltzmann Machine
The neurons, xi, are observable (that is, they represent the vectors that are generated by the process that the RBM has to learn), ...
