Abstract:Exploration seismology uses reflected and refracted seismic waves, emitted from a controlled (active) source into the ground, and recorded by an array of seismic sensors (receivers) to image the subsurface geological structures. These seismic images are the main resources for energy and resource exploration and scientific investigation of the crust and upper mantle. The exponential growth of data volumes from seismic surveys using modern instruments drives the need for algorithmic developments to process large-volume high-dimensional data. This is where artificial intelligence and modern machine learning techniques could play a key role. We survey recent advances in applications of machine learning (ML) methods, more specifically deep neural networks (DNNs), in the exploration seismology. We provide comprehensive and technically oriented review of DNNs applications for seismic data acquisition, data pre-processing tasks such as interpolation/extrapolation, denoising, first break picking, velocity picking, seismic migration, data processing tasks such as geological and structural interpretations, and data modeling tasks such as the inference of sub-surface structures and lithologic and petrophysical properties. Our goal is to document the technological trends, progress, and challenges for DNN approaches in the exploration seismology. This could benefit other sectors of the seismological community such as earthquake and engineering seismology by sharing how progress was achieved. data pre-processing tasks such as interpolation/extrapolation, denoising, first break picking, velocity picking, seismic migration, data processing tasks such as geological and structural interpretations, and data modeling tasks such as the inference of sub-surface structures and lithologic and petrophysical properties. Our goal is to document the technological trends, progress, and challenges for DNN approaches in the exploration seismology. This could benefit other sectors of the seismological com-#xD;munity such as earthquake and engineering seismology by sharing how progress was achieved.

Example Forgetting: A Novel Approach to Explain and Interpret Deep Neural Networks in Seismic Interpretation

Understanding Neural Networks through Representation Erasure.

Explaining How Deep Neural Networks Forget by Deep Visualization

Visualizing and Understanding Deep Neural Networks : A Case of Seismic Fault Detection Network

Automatic classification of geologic units in seismic images using partially interpreted examples

Example forgetting and rehearsal in continual learning

On Local Overfitting and Forgetting in Deep Neural Networks

Disentangling Noise Patterns from Seismic Images: Noise Reduction and Style Transfer

Measuring Catastrophic Forgetting in Neural Networks

Does shallow geological knowledge help neural-networks to predict deep units?

Overcoming Catastrophic Forgetting by XAI

Overcoming catastrophic forgetting in neural networks

Shape Carving Methods of Geologic Body Interpretation from Seismic Data Based on Deep Learning

SynapNet: A Complementary Learning System Inspired Algorithm With Real-Time Application in Multimodal Perception

Anatomy of Catastrophic Forgetting: Hidden Representations and Task Semantics

Dissecting Catastrophic Forgetting in Continual Learning by Deep Visualization

Applications of Deep Neural Networks in Exploration Seismology: a technical survey

Ensemble Deep Learning for enhanced seismic data reconstruction

Negotiated Representations to Prevent Forgetting in Machine Learning Applications

Fortuitous Forgetting in Connectionist Networks

Semantic Segmentation of Seismic Images