Comprehensive Study of Deep Learning

Abstract

Deep learning is the sub domain of machine learning with the representation learning capability to deal to learn representations of data with multiple levels of abstraction. It provide high computation rate to compose multiple processing layers. Deep learning is about the learning multiple levels of representation of data like image, speech etc. And abstraction means hide the unnecessary detail. Deep learning can identify complex structure easily applied back propagation algorithm on it to specify how a machine can change its internal parameter from the representation.