Class of deep learning models that learn to associate lower energy levels with more probable configurations of the input data.
Generality: 625
