Seismic waves from earthquakes and other sources are used to infer the structure and properties of Earth’s interior. The availability of large-scale seismic datasets and the suitability of deep-learning techniques for seismic data processing have pushed deep learning to the forefront of fundamental, long-standing research investigations in seismology. However, some aspects of applying deep learning to seismology are likely to prove instructive for the geosciences, and perhaps other research areas more broadly. Deep learning is a powerful approach, but there are subtleties and nuances in its application. We present a systematic overview of trends, challenges, and opportunities in applications of deep-learning methods in seismology. Description Large-scale learning The large amount and availability of datasets in seismology creates a great opportunity to apply machine learning and artificial intelligence to data processing. Mousavi and Beroza provide a comprehensive review of the deep-learning techniques being applied to seismic datasets, covering approaches, limitations, and opportunities. The trends in data processing and analysis can be instructive for geoscience and other research areas more broadly. —BG The ways in which deep learning can help process and analyze large seismological datasets are reviewed. BACKGROUND Seismology is the study of seismic waves to understand their origin—most obviously, sudden fault slip in earthquakes, but also explosions, volcanic eruptions, glaciers, landslides, ocean waves, vehicular traffic, aircraft, trains, wind, air guns, and thunderstorms, for example. Seismology uses those same waves to infer the structure and properties of planetary interiors. Because sources can generate waves at any time, seismic ground motion is recorded continuously, at typical sampling rates of 100 points per second, for three components of motion, and on arrays that can include thousands of sensors. Although seismology is clearly a data-rich science, it often is a data-driven science as well, with new phenomena and unexpected behavior discovered with regularity. And for at least some tasks, the careful and painstaking work of seismic analysts over decades and around the world has also made seismology a data label–rich science. This facet makes it fertile ground for deep learning, which has entered almost every subfield of seismology and outperforms classical approaches, often dramatically, for many seismological tasks. ADVANCES Seismic wave identification and onset-time, first-break determination for seismic P and S waves within continuous seismic data are foundational to seismology and are particularly well suited to deep learning because of the availability of massive, labeled datasets. It has received particularly close attention, and that has led, for example, to the development of deep learning–based earthquake catalogs that can feature more than an order of magnitude more events than are present in conventional catalogs. Deep learning has shown the ability to outperform classical approaches for other important seismological tasks as well, including the discrimination of earthquakes from explosions and other sources, separation of seismic signals from background noise, seismic image processing and interpretation, and Earth model inversion. OUTLOOK The development of increasingly cost-effective sensors and emerging ground-motion sensing technologies, such as fiber optic cable and accelerometers in smart devices, portend a continuing acceleration of seismological data volumes, so that deep learning is likely to become essential to seismology’s future. Deep learning’s nonlinear mapping ability, sequential data modeling, automatic feature extraction, dimensionality reduction, and reparameterization are all advantageous for processing high-dimensional seismic data, particularly because those data are noisy and, from the point of view of mathematical inference, incomplete. Deep learning for scientific discovery and direct extraction of insight into seismological processes is clearly just getting started. Aspects of seismology pose interesting additional challenges for deep learning. Many of the most important problems in earthquake seismology—such as earthquake forecasting, ground motion prediction, and rapid earthquake alerting—concern large and damaging earthquakes that are (fortunately) rare. That rarity poses a fundamental challenge for the data-hungry methods of deep learning: How can we train reliable models, and how do we validate them well enough to rely on them when data are scarce and opportunities to test models are infrequent? Further, how can we operationalize deep-learning techniques in such a situation, when the mechanisms by which they make predictions from data may not be easily explained, and the consequences of incorrect models are high? Incorporating domain knowledge through physics-based and explainable deep learning and setting up standard benchmarking and evaluation protocols will help ensure progress, as is the nascent emergence of a seismological data science ecosystem. More generally, a combination of data science literacy for geoscientists as well as recruiting data science expertise will help to ensure that deep-learning seismology reaches its full potential. Deep-learning processing of seismic data and incorporation of domain knowledge can lead to new capabilities and new insights across seismology. A. MASTIN/SCIENCE, TOP RIGHT: CARA HARWOOD/UNIVERSITY OF CALIFORNIA-DAVIS/CC BY-NC-SA 3.0 MIDDLE RIGHT: JOHAN SWANEPOEL/SCIENCE SOURCE

Deep-learning seismology

In this letter, the possibility of using device-free sensing (DFS) technology for personnel detection in a foliage environment is investigated. Although the conventional algorithm that based on statistical properties of the received-signal strength (RSS) for target detection at indoor or open-field environment has come a long way in recent years, it is still questionable if this algorithm is fully functional at outdoor with the changing atmosphere and ground conditions, such as a foliage environment. To answer this question, a variety of the measured data have been taken using different targets in a foliage environment. Applying these data along with support vector machine, the impact on detection accuracy due to different classification algorithms is studied. An algorithm that based on the extraction of the high-order cumulant (HOC) of the signals is presented, while the conventional RSS-based one is used as a benchmark. The measurement results show that the classification accuracy of the HOC-based algorithm is better than the RSS-based one by at least 17%. Moreover, to ensure the reliability of the HOC-based approach, the impact on classification accuracy due to different numbers of training samples and different values of signal-to-noise ratio is extensively verified using experimentally recorded samples. To the best of our knowledge, this is the first time that a DFS-based sensing approach is demonstrated to have a potential to distinguish between human and small-animal targets in a foliage environment.

Device-Free Sensing for Personnel Detection in a Foliage Environment

Millimeter-wave (mm-wave) monolithic-microwave-integrated-circuit (MMIC) bandpass filters (BPFs) usually feature high insertion loss (IL). To this concern, an analytical method for low-loss MMIC BPFs is presented by using lumped-distributed parameters in this article. To reduce the IL, the enhanced-quality-factor resonator and low-loss coupling topology are both investigated. First, the proposed resonator and filter are realized by combining quasi-lumped elements and distributed mutual coupling. Both compact size and multiple transmission zeroes are accordingly realized. Second, a low-loss coupling topology is further introduced for MMIC filter by using mixed electric and magnetic couplings. For theoretical analysis, the lossy equivalent circuit modeling for mm-wave MMIC BPF is investigated, as well as an analytical parameter extraction. Based on the circuit analysis, a fast synthesis of low-loss mm-wave MMIC filter is presented. For demonstration, two compact MMIC BPFs are designed by using the gallium arsenide (GaAs) process. Low IL of only 0.95 dB, sharp rejection skirt of passband, and wide stopband are simultaneously obtained. The features of the proposed MMIC BPFs indicate promising applications in the fifth-generation (5G) mm-wave systems.

Low Insertion-Loss MMIC Bandpass Filter Using Lumped-Distributed Parameters for 5G Millimeter-Wave Application

The microtremor survey method (MSM) has the potential to be an important geophysical method for identifying the strata velocity structure and detecting the buried fault structures. However, the existing microtremor exploration equipment has been unable to satisfy the requirements of the MSM, which suffers from low data accuracy, long measurement time, and blind acquisition. In this study, we combined a 2 Hz moving coil geophone with advanced acquisition systems to develop a new integrated energy-efficient wireless sensor node for microtremor exploration. A high-precision AD chip and noise matching technology are used to develop a low-noise design for the sensor node. Dynamic frequency selection technology (DFS) and dynamic power management technology (DPM) are used to design an energy-efficient mode. The data quality monitoring system solves the closed technical flaws between the acquisition systems and the control center via 4G wireless monitoring technologies. According to the results of a series of in situ tests and field measurements, the noise level of the system was 0.7 μV@500 Hz with 0 dB attenuation and 220 mW power consumption of the system in the autonomous data acquisition mode. Therefore, it provides substantial support for the effective data acquisition over long measurement durations in microtremor exploration processes. The applicability of the system is evaluated using field data, according to which the integrated energy-efficient wireless sensor node is convenient and effective for MSM.

https://www.mdpi.com/1424-8220/19/3/544/pdf

An Integrated Energy-Efficient Wireless Sensor Node for the Microtremor Survey Method.

This work presents radio frequency (RF) microelectromechanical system (MEMS)-based compact, high power, and reliable tunable bandpass filter for millimeter-wave RF front end. The design is fabricated on 635- $\mu \text{m}$ alumina substrate using surface micromachining process. All functional building blocks of the filter are fabricated and tested independently to ensure the optimum filter performances. Total area of the filter is only 2.3 mm2 including bias lines and pads. The proposed filter can be tuned to any frequency between 27 and 29 GHz and it demonstrates insertion loss of 1.92 dB and matching of 23.4 dB at 28 GHz with fractional bandwidth (FBW) of 6.7%. Filter also gives bandwidth tuning of 4.4 GHz with acceptable loss. Filter works satisfactorily up to 1 billion cycles with 1 W of power. Device reliability tested within a low-cost package. To the best of our knowledge, this is the first reported tunable MEMS bandpass filter in the literature which has undergone and presented with extensive reliability results.

Reliable, Compact, and Tunable MEMS Bandpass Filter Using Arrays of Series and Shunt Bridges for 28-GHz 5G Applications

Advanced micromachining technology has made magnificent progress for fabrication of non-planar circuits. Using this technology, circuits and systems can be implemented in a more cost-effective way. Unlike the conventional planar circuit, low-cost and highly sensitive misalignment sensor is required to detect imperfect placement of different micro-devices, which may be of the order of sub-micrometers. Currently, this is hardly to be achieved by using the existing approaches. In this letter, we present a novel sensor design approach utilizing the parasitic capacitance of an integrated coupled-line resonator for misalignment sensing. Due to vertical misalignment between two metal strips, the parasitic capacitance of the sensor varies, which results in a resonance shift from 53 to 68 GHz, while a reasonably strong transmission notch is still maintained. Taking advantage of this principle, misalignment can be effectively detected. To prove the concept, several devices are fabricated in a standard silicon technology. Three samples with the same structure are used to evaluate the reliability, while eight different structures are used to verify the concept. All results are extensively validated through both simulation and measurements.

Design of an On-Chip Highly Sensitive Misalignment Sensor in Silicon Technology

Recently, large scale seismic data acquisition has been a critical method for scientific research and industrial production. However, due to the bottleneck on data transmission and the limitation of energy storage, it is hard to conduct large seismic data acquisition in a real-time way. So, in this paper, an efficient seismic data acquisition method, namely, compressed sensing architecture with generative adversarial networks (CSA-GAN), is proposed to tackle the two restrictions of collecting large scale seismic data. In the CSA-GAN, a data collection architecture based on compressed sensing theory is applied to reduce data traffic load of the whole system, as well as balance the data transmission. To make the compressed sensing procedure perform well in both data quality and compression ratio, a kind of generative adversarial networks is designed to learn the recovering map. According to our experiment results, a high data quality (about 30 dB) at the compression ratio of 16 is achieved by the proposed approach, which enables the system to afford 15 times more sensors and reduces the energy cost by means of data collection from N(N + 1)/2 to N
 2
/16. These results show that the CSA-GAN can afford more sensors with the same bandwidth and consume less energy, via improving the efficiency seismic data acquisition.

/pdf/an-efficient-seismic-data-acquisition-based-on-compressed-4c3ut8dmme.pdf

An Efficient Seismic Data Acquisition Based on Compressed Sensing Architecture With Generative Adversarial Networks

Many deep learning methods have been proposed to classify moving targets from seismic signals in recent years. However, the existing deep models are all designed based on the ``end-cloud'' framework, in which real-time data processing is difficult because of communication delays. To address this problem and achieve on-site target classification, we propose a novel edge intelligence-oriented method, named compressed sensing-edge convolutional neural network (CS-ECNN). In this method, the acquired seismic signals are first mapped onto a compressed domain using CS. This operation reduces data dimensions, while being able to retain the vast majority of valuable seismic features. Following that, a convolutional neural network is employed to extract implicit features directly from the compressed seismic measurements and then classify the feature vectors. To evaluate the proposed method, the seismic data recorded in DARPA's SensIT project are used as a case study. The experimental results demonstrate that the proposed model is edge-matched, and it achieves comparable classification accuracy to the state-of-the-art cloud-based models with only 1/10 computation time.

Edge Intelligence-Based Moving Target Classification Using Compressed Seismic Measurements and Convolutional Neural Networks

In modern seismic data acquisition, real-time data collection is a challenging task due to bandwidth limitations in wireless communications. In this letter, we propose a novel compressive data gathering scheme using generative adversarial networks, named GAN-CDG, to improve the efficiency of data gathering. Instead of collecting the originally acquired data, GAN-CDG gathers data projections in wireless geophone networks. Data compression and load-balanced relay transmission are utilized during the projection process. To speed up the formation of projections, the shortest path routing tree (SPRT) is constructed, which achieves the minimum end-to-end time delay. The sparse domain of seismic signals and its reconstruction mapping are learned by sparsity-constrained adversarial networks. The testing results demonstrate that projections with high compression ratios (e.g., 16) are gathered efficiently with the SPRT. Then, original seismic signals can be reconstructed accurately (over 30 dB) from the projections using the adversarial model, which outperforms the state-of-the-art method.

Xunqian Tong

Papers

Design of an On-Chip Highly Sensitive Misalignment Sensor in Silicon Technology

An Efficient Seismic Data Acquisition Based on Compressed Sensing Architecture With Generative Adversarial Networks

Edge Intelligence-Based Moving Target Classification Using Compressed Seismic Measurements and Convolutional Neural Networks

Compressive Data Gathering With Generative Adversarial Networks for Wireless Geophone Networks

Moving target recognition with seismic sensing: A review