Deep Learning is part of a family of machine learning techniques based on learning representations of data. Deep Learning is loosely based on our brain's own neural networks, the purpose of this structure is to provide a large number of highly interconnected elements (in biological systems, this would be the neurons in our brains); there are approximately 100 billion neurons in our brain, each connected to approximately 10,000 other neurons, resulting in a mind-boggling 1015 synaptic connections. These elements work together to solve problems through learning processes – examples include pattern recognition and data classification.
Learning within this architecture involves modifications of the connections between the interconnected elements similar to how our own brains make adjustments to the synaptic connections between neurons:
Source: Wikimedia Commons: File: Réseau de neurones.jpg; https://commons.wikimedia.org/wiki/File:Réseau_de_neurones.jpg.