Since Professor Hinton et al. using the deep learning model based on convolutional neural networks won the ImageNet Classification Contest, deep learning sweeps across various industries. By introducing the history of deep learning, this paper explores the use of the deep learning model in geological industry of China. It also presents the basic concepts, such as neuron, neural network, supervised learning and unsupervised learning. Based on those concepts, two important networks of Deep Belief Networks (DBN) and Convolutional Neural Networks (CNN) are introduced. In the end, with reference to the application in medicine, it puts forward the application prospect of deep learning in geology. First, deep learning has a great advantage in computer vision, which can be used in remote sensing image clustering, rock sample classification and rock sheet data description. Second, its precise identification of original data could help identify geologic anomaly data to determine the possible location of ore-forming targets. Third, deep learning is of great help in sound signal data processing before earthquake, which can determine the remaining time before the earthquake occurred. |