Abstract

Deep learning is fast emerging as a potential disruptive tool to tackle longstanding research problems across the sciences. This is mainly driven by its ability to find complex patterns in large datasets without the need for feature extraction or engineering. It comes as no surprise that geophysicists, like domain experts from other scientific disciplines, are starting to find value in deep learning methods. Numerous articles demonstrating the efficacy of deep learning to longstanding geophysical problems have appeared in the recent past. These include problems such as earthquake detection, ground-motion prediction, seismic tomography, and earthquake early warning among many others. However, there is a growing concern among some practitioners on the overuse of deep learning in the sciences. Since these deep learning algorithms are so complicated, it is often impossible to reason about exactly how the inputs have been manipulated. As these algorithms begin to be applied widely, the risk of misinterpretations and erroneous conclusions may spiral. In this talk, we will explore deep learning solutions to a multitude of geophysical problems including detection and location of passive seismic events and solving non-linear partial differential equations, such as the eikonal equation and the Helmholtz equation. Through these examples, we will analyze the opportunities and pitfalls associated with deep learning applied to geophysical exploration.

---