Abstract: Deep learning is a machine learning algorithm which learns intrinsic representations of complex data by multi-layer abstractions of information. It has dramatically improved the state-of-the-art in visual object recognition and detection, speech recognition, and many other artificial intelligence computing tasks. This paper will first introduce the basics of the algorithm and the motivations of its application in high energy physics computing. The applications of deep neural, convolutional neural and recursive neural networks as well as other models of deep learning in high energy physics will be summarized, together with several case studies. Finally, problems and potential solutions in the integration of deep learning with current high energy computing environments will be briefly mentioned.