A generative model made by stacking several RBMs in layers.
A DBN is like a school lunch line for data. Each station adds a clue, until the last one can guess the whole meal.
It was a big early spark for Deep Learning. People used it to learn features and make new examples.
RBM
A DBN often stacks several RBMs in layers.
Deep Learning
DBN layer-by-layer pretraining helped revive Deep Learning.
Generative Model
A DBN can learn data patterns and make new samples.
Unsupervised Learning
Early DBNs often used Unsupervised Learning to pretrain each layer.