The estimation of sparse shallow-water acoustic communication channels and the impact performance on equalization phase coherent signals are investigated. Given sufficiently wide transmission bandwidth, impulse response channel is often as multipath arrivals become resolvable. In presence significant surface waves, associated with scattering fluctuate rapidly over time, in sense that complex gain, arrival Dopplers each all change dynamically. A technique developed based delay-Doppler-spread...

Mapping of seismic and lithologic facies from 3D reflection data plays a key role in depositional environment analysis reservoir characterization during hydrocarbon exploration development. Although variety machine-learning methods have been developed to speed up interpretation improve prediction accuracy, there still exist significant challenges multiclass classification practice. Some these limitations include complex representation, limited training with labels, imbalanced class...

Machine learning, and specifically deep-learning (DL) techniques applied to geophysical inverse problems, is an attractive subject, which has promising potential and, at the same time, presents some challenges in practical implementation. Some obstacles relate scarce knowledge of searched geologic structures, a problem that can limit interpretability generalizability trained DL networks when independent scenarios real applications. Commonly used (physics-driven) least-squares optimization...

Recent developments in industrial systems provide us with a large amount of time series data from sensors, logs, system settings and physical measurements, etc. These are extremely valuable for providing insights about the complex could be used to detect anomalies at early stages. However, special characteristics these data, such as high dimensions dependencies between variables, well its massive volume, pose great challenges existing anomaly detection algorithms. In this paper, we propose...

Patenting is one of the most important ways to protect company's core business concepts and proprietary technologies. Analyzing large volume patent data can uncover potential competitive or collaborative relations among companies in certain areas, which provide valuable information develop strategies for intellectual property (IP), R&D, marketing. In this paper, we present a novel topic-driven analysis mining system. Instead merely searching over content, focus on studying heterogeneous...

A rapid focus-detection technique based directly on the spectral content of digital holograms is developed. It differs from previous approaches in that it does not need a full reconstruction image. The uses l1 norms object components associated with real and imaginary parts kernel. Further, can be computed efficiently spatial frequency domain using polar coordinate system, yielding drastic speedup ∼2 orders magnitude compared image-based focus detection. Significant computational savings are...

We have developed a method to combine unsupervised and supervised deep-learning approaches for seismic ground roll attenuation. The consists of three components that physical meaning motivation. first component is convolutional neural network (CNN) separate record into signal, while minimizing the residual between sum generated signal from two subnetworks input record. second creates maximum separation in f- k domain, by training classifier. third CNN mapping roll, which overcomes problem...

Tensor decomposition is a fundamental tool for analyzing multi-dimensional data by learning low-rank factors to represent high-order interactions. While recent works on temporal tensor have made significant progress incorporating continuous timestamps in latent factors, they still struggle with general indexes not only the mode but also other modes, such as spatial coordinates climate data. Additionally, problem of determining rank remains largely unexplored models. To address these...

The equations to reconstruct an image plane from a hologram are developed. This development is carried out for planes parallel the hologram, which allows fast computation through use of Fourier transforms. Algorithms digital computer developed so images can be reconstructed, both with and without Fresnel approximation, digitized need three-dimensional optical reconstruction equipment. Examples holographically recorded marine micro-organisms shown. A computational method counting number in...

A typical assumption in supervised fault detection is that abundant historical data are available prior to model learning, where all types of faults have already been observed at least once. This likely be violated practical settings as new can emerge over time. In this paper we study often overlooked cold start learning problem data-driven detection, the beginning only normal operation and faulty become occur. We explored how leverage strengths unsupervised approaches build a capable...

Reservoir characterization and monitoring represent some of the most ambitious goals for geophysical methods. Several challenges are involved, including sensitivity to parameter changes resolution obtained results. Electromagnetic (EM) methods attractive reservoir applications due high resistivity oil/water saturations. Crosswell EM surface-to-borehole provide opportunities monitoring. The inverse problem, however, is highly nonconvex ill-posed so as necessitate significant preconditioning...

Compressional and shear sonic traveltime logs (DTC DTS, respectively) are crucial for subsurface characterization seismic-well tie. However, these two often missing or incomplete in many oil gas wells. Therefore, petrophysical geophysical workflows include log synthetization pseudo-log generation based on multivariate regression rock physics relations. Started March 1, 2020, concluded May 7, the SPWLA PDDA SIG hosted a contest aiming to predict DTC DTS from seven “easy-to-acquire”...

We develop a novel physics-adaptive machine-learning (ML) inversion scheme showing optimal generalization capabilities for field data applications. apply the physics-driven deep-learning to massive helicopter-borne transient electromagnetic (TEM) set. The objective is accurate modeling of near surface enhancing exploration low-relief structures in sand covered desertic area. Enhanced neural network (NN) or other ML techniques obtained from automatic physics-based adaptive training using...

Automated detection and classification of geological structure elements, such as salt domes, channels, faults folds, from seismic images provides an important first step towards new generation interpretation tools. Meaningful structural hypothesis can later be constructed these elements their configurations. Much work has been developed in the past this area, yet significant challenges still remain for automated analysis with adequate efficiency accuracy, including lack high quality...

We experimentally demonstrated the contrast enhancement in a Ti:sapphire chirped-pulse amplification (CPA) laser with noncollinear femtosecond optical-parametric amplifier. A total gain of 3.4 × 10(4) and pulse energy 26 μJ were achieved. With clean high-energy seeding pulse, ratio main amplified to spontaneous emission Ti: sapphire CPA system was improved around 10(10) within time scale hundreds picoseconds.

Machine learning (ML) and specifically deep (DL) techniques applied to inversion problems are still a relatively new area of research which is appealing geophysical applications. We developed hybrid workflow combining the efficiency physics-driven with power data-driven DL based inversion. The two procedures coupled by model term. method involves re-training network after each iterations. schemes evolving balancing other converge common satisfying data misfit criteria optimization parameters...

Distributed acoustic sensing (DAS) has emerged as a transformational technology for seismic data acquisition. However, noise remains major impediment, necessitating advanced denoising techniques. This study pioneers the application of diffusion models, type generative model, DAS vertical profile (VSP) denoising. The network is trained on new generated synthetic dataset that accommodates variations in acquisition parameters. model applied to suppress and field DAS-VSP data. results...

Depth matching or depth shifting between well logs acquired from different runs core scans and is a critical data quality control task to ensure subsequent accurate petrophysical interpretation modeling. Conventional depth-shifting workflow heavily relies on human expertise manually match series of peaks troughs log curves, which often subjective, error-prone, cumbersome. Therefore, it necessary establish an automatic perform this routine yet important accurately in consistent efficient...

Seismic facies analysis interprets depositional environment and types from the reflection seismic data, an important step in exploration reservoir characterization. While machine learning methods, especially deep models such as convolutional neural networks (CNNs) have been applied to assist interpretation salt identification, significant challenges still remain for 3D multi-class classification: complex data representation, limited labeled training, imbalanced class distribution lack of...

One important indicator of the quality a hands-on measure job performance is its dependability. Due to complexity such measurements, two concerns arise: (a) Can test examiners reliably score performance, and (b) can complexities measurements be modeled statistically evaluate reliability tests? Generalizability theory (G theory) was used model U.S. Marine Corps infantry riflemen. Results show that experienced, well- trained, calibrated dependably performance; several ways examining...

Patents are critical for a company to protect its core technologies. Effective patent mining in massive databases can provide companies with valuable insights develop strategies IP management and marketing. In this paper, we study novel problem of automatically discovering patents (i.e., high novelty influence domain). We address the unique vocabulary usage problem, which is not considered traditional word-based statistical methods, propose topic-based temporal approach quantify patent's...