Well to explain deep learning you require understanding of basic ai and machine learning or neural network in the first place.

At the most simplest term is that you use some algorithms to interpret some sort of data (interpretation is being used loosely here). Then when you connect and layer these algorithms and each layer might provide different interpretation then you have a deep learning system that is able to automatically (also loose definition in term of supervision) learns from unstructured and unlabeled data.

Hope its not Greek to you, but IT/IS people always sensitive in the term used as it might lead to different meaning.