A standard BM network is divided into a set of observable and visible units $\boldsymbol{x}$ and a set of unknown hidden units/nodes $\boldsymbol{h}$.

Additionally there can be bias nodes for the hidden and visible layers. These biases are normally set to $1$.

BMs are stackable, meaning they cwe can train a BM which serves as input to another BM. We can construct deep networks for learning complex PDFs. The layers can be trained one after another, a feature which makes them popular in deep learning