Deep Learning Realm for Geophysics: Seismic Acquisition, Processing, Interpretation, and Inversion

Abstract

Applying deep-learning models to geophysical applications has attracted special attentions during the past a couple of years. There are several papers published in this domain involving with different topics primarily focusing on synthetic data. Based on the presented results, reaching the satisfaction from geophysicists is largely far away the performance of the conventional algorithms. Moreover, it is extremely hard to find documents with detailed illustrations about what are the best practices that researchers should follow regarding how to design appropriate deep-learning models to effectively and precisely tackle problems relevant to the field data. This paper serves as a summarization to demonstrate successful stories and share with extensive experiences we have gained during past several years in the process of designing and deploying of deep-learning models to the geophysical projects in a large scale. Four different disciplines are discussed individually with seismic acquisition, processing, interpretation, and inversion. Finally, special attentions about designing an effective deep-neural-networks especially for geophysics are discussed.