Deep learning application in Earthquake related studies in Indonesia: A review Lina Handayani
Indonesian Institute of Sciences
Abstract
The basic of deep learning in earth sciences is statistics and geophysics. And recently, the studies in this area have been increasing, for the availability of a high volume of open access data. DL is applicable in all geophysical methods, which is one of the approaches to simplify or to accelerate some processes, or to have more elaborate results. The application for ^artificial intelligence^ approaches in earthquakes has been used since the computation was available. Two or three decades ago, we approach the problem with fuzzy logic and neural networks. Recently, we used deep learning, which has at least three layers of neural networks. In this paper, we review the advance of the usability in several Earthquake related studies. Earthquake prediction is one of the most desired results. With advanced methods and more data, we expected better results. However, decades of studies indicate small advancement. The closest outcome is forecasting the aftershocks. However, some studies offered good results in speedy and more accurate seismic arrival waves processing. In addition, one applicable ready result is the study of hazard and risk assessment.
