Deep Learning refers to complex multi-layer neural nets.  They are especially suitable for image and voice recognition, and for unsupervised tasks with complex, unstructured data.