Considering a Boltzmann machine as a set of units $x$ with a weight matrix $U$ and biases $b$, the standard energy function is given by: $E(x) = -x^{T}Ux-b^{T}x$. We can implement structures of Boltzmann machines with different numbers of layers by partitioning $x$ into visible and hidden units and zeroing out elements of $U$ for units that do not interact. A centered deep Boltzmann machine reparametrizes this energy function as: E^{'}(x;U,b) = -(x-mu)^{T}U(x-mu)-(x-mu)^{T}b. Typically, $\mu$ is a hyperparameter fixed at the beginning of training, and it is usually chosen to ensure that $x-\mu \approx 0$ when the model is initialized.