Understanding the principles of geophysical phenomena is an essential and challenging task. "Model-driven" approaches have supported the development of geophysics for a long time; however, such methods suffer from the curse of dimensionality and may inaccurately model the subsurface. "Data-driven" techniques may overcome these issues with increasingly available geophysical data. In this article, we review the basic concepts of and recent advances in data-driven approaches from dictionary learning to deep learning in a variety of geophysical scenarios. Explorational geophysics including data processing, inversion and interpretation will be mainly focused. Artificial intelligence applications on geoscience involving deep Earth, earthquake, water resource, atmospheric science, satellite remoe sensing and space sciences are also reviewed. We present a coding tutorial and a summary of tips for beginners and interested geophysical readers to rapidly explore deep learning. Some promising directions are provided for future research involving deep learning in geophysics, such as unsupervised learning, transfer learning, multimodal deep learning, federated learning, uncertainty estimation, and activate learning.
Seismic data quality is vital to geophysical applications, so methods of data recovery, including denoising and interpolation, are common initial steps in the seismic data processing flow. We present a method to perform simultaneous interpolation and denoising, which is based on double-sparsity dictionary learning. This extends previous work that was for denoising only. The original double sparsity dictionary learning algorithm is modified to track the traces with missing data by defining a masking operator that is integrated into the sparse representation of the dictionary. A weighted low-ran…
Modeling geophysical processes as low-dimensional dynamical systems and regressing their vector field from data is a promising approach for learning emulators of such systems. We show that when the kernel of these emulators is also learned from data (using kernel flows, a variant of cross-validation), then the resulting data-driven models are not only faster than equation-based models but are easier to train than neural networks such as the long short-term memory neural network. In addition, they are also more accurate and predictive than the latter. When trained on geophysical observational d…
Recent years have seen a surge in interest in building deep learning-based fully data-driven models for weather prediction. Such deep learning models if trained on observations can mitigate certain biases in current state-of-the-art weather models, some of which stem from inaccurate representation of subgrid-scale processes. However, these data-driven models, being over-parameterized, require a lot of training data which may not be available from reanalysis (observational data) products. Moreover, an accurate, noise-free, initial condition to start forecasting with a data-driven weather model …
Los Alamos National Laboratory（洛斯阿拉莫斯国家实验室）；Microsoft Research（微软研究院）；Lawrence Berkeley National Laboratory（劳伦斯伯克利国家实验室）
Multi-physical inversion plays a critical role in geophysics. It has been widely used to infer various physical properties~(such as velocity and conductivity). Among those inversion problems, some are explicitly governed by partial differential equations~(PDEs), while others are not. Without explicit governing equations, conventional multi-physical inversion techniques will not be feasible and data-driven inversion requires expensive full labels. To overcome this issue, we develop a new data-driven multi-physics inversion technique with extremely weak supervision. Our key finding is that the p…
University of Cambridge（剑桥大学）；Lancaster University（兰卡斯特大学）；British Antarctic Survey（英国南极调查局）
Ensembles of geophysical models improve projection accuracy and express uncertainties. We develop a novel data-driven ensembling strategy for combining geophysical models using Bayesian Neural Networks, which infers spatiotemporally varying model weights and bias while accounting for heteroscedastic uncertainties in the observations. This produces more accurate and uncertainty-aware projections without sacrificing interpretability. Applied to the prediction of total column ozone from an ensemble of 15 chemistry-climate models, we find that the Bayesian neural network ensemble (BayNNE) outperfo…
Autoencoders are neural network formulations where the input and output of the network are identical and the goal is to identify the hidden representation in the provided datasets. Generally, autoencoders project the data nonlinearly onto a lower dimensional hidden space, where the important features get highlighted and interpretation of the data becomes easier. Recent studies have shown that even in the presence of noise in the input data, autoencoders can be trained to reconstruct the noisefree component of the data from the reduced-dimensional hidden space. In this paper, we explore the app…
Self-Supervised Knowledge-Driven Deep Learning for 3D Magnetic Inversion
Beijing National Research Center for Information Science and Technology (BNRist), Department of Automation, Tsinghua University（清华大学自动化系北京国家信息科学技术研究中心）；School of Civil Engineering, Changsha University of Science and Technology（长沙理工大学土木工程学院）
Salk Institute for Biological Studies（索尔克生物研究所）；U.S. Department of the Interior（美国内政部）；Scripps Institution of Oceanography（斯克里普斯海洋研究所）；University of California San Diego（加州大学圣地亚哥分校）
Data-driven, model-free analytics are natural choices for discovery and forecasting of complex, nonlinear systems. Methods that operate in the system state-space require either an explicit multidimensional state-space, or, one approximated from available observations. Since observational data are frequently sampled with noise, it is possible that noise can corrupt the state-space representation degrading analytical performance. Here, we evaluate the synthesis of empirical mode decomposition with empirical dynamic modeling, which we term empirical mode modeling, to increase the information cont…
The geophysics of extrasolar planets is a scientific topic often regarded as standing largely beyond the reach of near-term observations. This reality in no way diminishes the central role of geophysical phenomena in shaping planetary outcomes, from formation, to thermal and chemical evolution, to numerous issues of surface and near-surface habitability. We emphasize that for a balanced understanding of extrasolar planets, it is important to look beyond the natural biases of current observing tools, and actively seek unique pathways to understand exoplanet interiors as best as possible during …
Using data alone, without knowledge of underlying physical models, nonlinear discrete time regional forecasting dynamical rules are constructed employing well tested methods from applied mathematics and nonlinear dynamics. Observations of environmental variables such as wind velocity, temperature, pressure, etc allow the development of forecasting rules that predict the future of these variables only. A regional set of observations with appropriate sensors allows one to forgo standard considerations of spatial resolution and uncertainties in the properties of detailed physical models. Present …
When monitoring geophysical parameters, data from segments that are contaminated by noise may have to be abandoned. This is the case, for example, in the geoelectrical field measurements at some sites in Japan, where high noise -due mainly to leakage currents from DC driven trains- prevails almost during 70\% of the 24 hour operational time. We show that even in such a case, the identification of seismic electric signals (SES), which are long-range correlated signals, may be possible, if the remaining noise free data are analyzed in natural time along with detrended fluctuation analysis (DFA).
