Search

Recent studies have demonstrated the potential of machine learning methods for fast and accurate mineral classification based on microscope thin sections. Such methods can be extremely useful to support geoscientists during the phases of operational geology, especially when mineralogical and petrological data are fully integrated with other geological and geophysical information. In order to be effective, these methods require robust machine learning models trained on pre-labeled data. Furthermore, it is mandatory to optimize the hyper-parameters of the machine learning techniques in order to guarantee optimal classification accuracy and reliability. Nowadays, deep learning algorithms are widely applied for image analysis and automatic classification in a large range of Earth disciplines, including mineralogy, petrography, paleontology, well-log analysis, geophysical imaging, and so forth. The main reason for the recognized effectiveness of deep learning algorithms for image analysis is that they are able to quickly learn complex representations of images and patterns within them. Differently from traditional image-processing techniques based on handcrafted features, deep learning models automatically learn and extract features from the data, capturing, in almost real-time, complex relationships and patterns that are difficult to manually define. Many different types of deep learning models can be used for image analysis and classification, including fully connected deep neural networks (FCNNs), convolutional neural networks (CNNs or ConvNet), and residual networks (ResNets). In this paper, we compare some of these techniques and verify their effectiveness on the same dataset of mineralogical thin sections. We show that the different deep learning methods are all effective techniques in recognizing and classifying mineral images directly in the field, with ResNets outperforming the other techniques in terms of accuracy and precision. In addition, we compare the performance of deep learning techniques with different machine learning algorithms, including random forest, naive Bayes, adaptive boosting, support vector machine, and decision tree. Using quantitative performance indexes as well as confusion matrixes, we demonstrate that deep neural networks show generally better classification performances than the other approaches. Furthermore, we briefly discuss how to expand the same workflow to other types of images and geo-data, showing how this deep learning approach can be generalized to a multiscale/multipurpose methodology addressed to the analysis and automatic classification of multidisciplinary information. This article has tutorial purposes, too. For that reason, we will explain, with a didactical level of detail, all the key steps of the workflow. [ABSTRACT FROM AUTHOR]

Analyzing geological drilling hole images acquired by Axial View Panoramic Borehole Televiewer (APBT) is a key step to explore the geological structure in a geological exploration. Conventionally, the borehole images are examined by technicians, which is inefficient and subjective. In this paper, three dominant types of borehole-wall images on coal-rock mass structure, namely, border images, fracture images and intact rock mass images are mainly studied. The traditional image classification methods based on unified feature extraction algorithm and single classifier is not effect for the borehole images. Therefore, this paper proposes a novel two-stage classification approach to improve the classification performance of borehole images. In the first-stage classification, the border images are identified from three kinds of images based on texture features and gray-scale histograms features. For the remaining two types of images, in the second-stage classification, Gabor filter is first applied to segment the region of interest (ROI) (such as microfracture, absciss layer and horizontal cracks, etc.) and the central interference region. Then, using the same feature vector after eliminating the central interference region, fracture images are separated from intact rock mass images. We test our two-stage classification system with real borehole images. The results of experimental show that the two-stage classification method can effectively classify three major borehole-wall images with the correction rate of 95.55% in the first stage and 95% in the second stage. [ABSTRACT FROM AUTHOR]

Rock types, pore structures and permeability are essential petrophysical outputs, and they contribute considerably to the highest degree of uncertainty in reservoir characterisation. These factors have a strong influence on exploration and field development decisions. Core analysis is the best approach for estimating permeability, assigning rock types and characterising pore networks. Wireline logs are the most often employed method for estimating the parameters at each data point of reservoirs since there are more un-cored wells than cored wells. Artificial intelligence, on the other hand, is gaining popularity in the geosciences due to the ever-increasing complexity and volume of available subsurface data. This is also obvious in the demand for faster and more accurate interpretations in order to identify reservoir characteristics in increasingly difficult and complicated petroliferous basins. Artificial Neural Networks and Self-Organizing Maps are examples of machine learning approaches that can be used in both supervised and unsupervised modes for modelling and prediction. Eocene carbonates of Mukta oilfield are the major pay rocks of strong geological heterogeneity in terms of their porosity and permeability relationship with pore structures. This paper outlines a novel method of rock fabric classification, pore structure characterization, flow unit classification and robust reservoir permeability modelling based on an integrated approach that incorporates core measurements, log data and machine learning techniques. The pore structure has been characterised by the combination of conventional core, capillary pressure and nuclear magnetic resonance data. Artificial neural network has added an adequate benefit in accurate permeability modelling by utilizing the concepts of rock classifications and hydraulic flow units. [ABSTRACT FROM AUTHOR]

This paper describes a method for training neural networks to learn circular dependencies. Variables with circular structure (e.g., time of day, day of year, and Earth location) appear in many different contexts within geoscience and remote sensing. Some common representations of circular variables (e.g., time of day in hours) can introduce discontinuities or topological distortions in estimation problems. They do not necessarily prevent a neural network from learning a relationship with circular dependencies. However, using topologically appropriate representations of circular variables can reduce the complexity necessary for a neural network to accurately learn such a relationship despite possibly increasing the number of inputs thereby reducing training times. In this paper, neural networks are trained to learn fictitious geophysical functions of time of day, of geolocation, and of time of year. In all three examples, using topologically appropriate representations of time and geolocation instead of conventional representations as inputs significantly reduced rms errors. Neural networks are also trained to learn the variations of air temperature observations with time of day and day of year, and one-month averages of sea surface temperature with geolocation. The improvement achieved by using topologically appropriate representations was limited by a natural random behavior in the data. However, there is a small but significant improvement in estimating the sea surface temperatures. Using the topologically appropriate representations of circular data when training the neural networks could be important in global Earth science remote sensing contexts, where significant diurnal, seasonal, or geographical variations exist. The studies presented also suggest the development of a more general framework for training neural networks that considers the topology of variables. [ABSTRACT FROM AUTHOR]

The fragile ecological environment near mines provide advantageous conditions for the development of landslides. Mine landslide susceptibility mapping is of great importance for mine geo-environment control and restoration planning. In this paper, a total of 493 landslides in Shangli County, China were collected through historical landslide inventory. 16 spectral, geomorphic and hydrological predictive factors, mainly derived from Landsat 8 imagery and Global Digital Elevation Model (ASTER GDEM), were prepared initially for landslide susceptibility assessment. Predictive capability of these factors was evaluated by using the value of variance inflation factor and information gain ratio. Three models, namely artificial neural network (ANN), support vector machine (SVM) and information value model (IVM), were applied to assess the mine landslide sensitivity. The receiver operating characteristic curve (ROC) and rank probability score were used to validate and compare the comprehensive predictive capabilities of three models involving uncertainty. Results showed that ANN model achieved higher prediction capability, proving its advantage of solve nonlinear and complex problems. Comparing the estimated landslide susceptibility map with the ground-truth one, the high-prone area tends to be located in the middle area with multiple fault distributions and the steeply sloped hill. [ABSTRACT FROM AUTHOR]

Building upon recent Deep Neural Network architectures, current approaches lying in the intersection of Computer Vision and Natural Language Processing have achieved unprecedented breakthroughs in tasks like automatic captioning or image retrieval. Most of these learning methods, though, rely on large training sets of images associated with human annotations that specifically describe the visual content. In this paper we propose to go a step further and explore the more complex cases where textual descriptions are loosely related to the images. We focus on the particular domain of news articles in which the textual content often expresses connotative and ambiguous relations that are only suggested but not directly inferred from images. We introduce an adaptive CNN architecture that shares most of the structure for multiple tasks including source detection, article illustration and geolocation of articles. Deep Canonical Correlation Analysis is deployed for article illustration, and a new loss function based on Great Circle Distance is proposed for geolocation. Furthermore, we present BreakingNews, a novel dataset with approximately 100K news articles including images, text and captions, and enriched with heterogeneous meta-data (such as GPS coordinates and user comments). We show this dataset to be appropriate to explore all aforementioned problems, for which we provide a baseline performance using various Deep Learning architectures, and different representations of the textual and visual features. We report very promising results and bring to light several limitations of current state-of-the-art in this kind of domain, which we hope will help spur progress in the field. [ABSTRACT FROM PUBLISHER]

Limb sounding instruments provide high vertical resolution data on the temperature and composition of the atmosphere. Their data are, therefore, valuable for assimilating into general circulation models of the atmosphere. Direct assimilation of radiances from limb sounders is more complex in practice than from nadir sounders due to the need to know the tangent pressures of the measurements. This paper discusses the practical implications of tangent pressures in direct radiance assimilation of limb sounding radiances and demonstrates that a neural network can be used to find these tangent pressures for the Earth Observing System Microwave Limb Sounder with a root mean-square error of σ = 50 m, which is comparable with that in traditional retrieval techniques. [ABSTRACT FROM AUTHOR]

The Shuttle Radar Topography Mission (SRTM) digital elevation model (DEM) is extensively utilized in various fields, such as forestry, oceanography, geology, and hydrology. However, due to limitations in radar side-view imaging, the SRTM DEM still contains gaps and anomalies, particularly in areas with an intricate topography, like forests. To enhance the accuracy of the SRTM DEM in forested regions, commonly employed approaches include regression analysis and artificial neural networks (ANN). Nevertheless, existing regression methods struggle to accurately capture the intricate nonlinear relationship between the error and influencing factors. Additionally, traditional ANN models are susceptible to overfitting, resulting in subpar accuracy. Deep Belief Network (DBN) is a highly precise algorithm in deep learning. However, the intricate combination of hyperparameters often leads to limited generalization ability and model robustness when correcting DEM. The present study proposes an error prediction model based on the DBN optimized by Particle Swarm Optimization (PSO) for SRTM DEM correction. By utilizing the PSO algorithm, we aim to identify the optimal combination of hyperparameters of DBN, including the number of neurons in the hidden layer and the learning rates. The experiment focuses on two regions in Hunan Province, China, characterized by abundant vegetation cover. The reference data utilized for comparison is ICESat/GLAS data. The experimental results demonstrate that the mean error (ME) and root mean square error (RMSE) of the SRTM DEM corrected by the proposed algorithm in these two regions are significantly reduced by 93.5%–96.0% and 21.5%–23.5%, respectively. Moreover, there is an improvement of over 26.1% in accuracy within complex terrain areas. Specifically, in broadleaf forest, the PSO-DBN method exhibits a remarkable accuracy improvement of 26.2%, while the DBN-corrected SRTM DEM shows an improvement of 15.3%. In coniferous forest, the PSO-DBN method achieves an accuracy improvement of 14.8%, whereas the DBN-corrected SRTM DEM demonstrates a gain of 5.8%. The approach provides a more effective and robust tool for correcting SRTM DEM or other similar DEMs over vegetated mountain areas. [ABSTRACT FROM AUTHOR]

Knowledge of tree species composition in a forest is an important topic in forest management. Accurate tree species maps allow for much more detailed and in-depth analysis of biophysical forest variables. The paper presents a comparison of three classification algorithms: support vector machines (SVM), random forest (RF) and artificial neural networks (ANN) for tree species classification using airborne hyperspectral data from the Airborne Prism EXperiment sensor. The aim of this paper is to evaluate the three nonparametric classification algorithms (SVM, RF and ANN) in an attempt to classify the five most common tree species of the Szklarska Poręba area: spruce (Picea alba L. Karst), larch (Larix decidua Mill.), alder (Alnus Mill), beech (Fagus sylvatica L.) and birch (Betula pendula Roth). To avoid human introduced biases a 0.632 bootstrap procedure was used during evaluation of each compared classifier. Of all compared classification results, ANN achieved the highest median overall classification accuracy (77%) followed by SVM with 68% and RF with 62%. Analysis of the stability of results concluded that RF and SVM had the lowest variance of overall accuracy and kappa coefficient (12 percentage points) while ANN had 15 percentage points variance in results.

Abstract: Geographic features have traditionally been visualized with fairly high amount of geometric detail, while relationships among these features in attribute space have been represented at a much coarser resolution. This limits our ability to understand complex high-dimensional relationships and structures existing in attribute space. In this paper, we present an alternative approach aimed at creating a high-resolution representation of geographic features with the help of a self-organizing map (SOM) consisting of a large number of neurons. In a proof-of-concept implementation, we spatialize 200,000+ U.S. Census block groups using a SOM consisting of 250,000 neurons. The geographic attributes considered in this study reflect a more holistic representation of geographic reality than in previous studies. The study includes 69 attributes regarding population statistics, land use/land cover, climate, geology, topography, and soils. This diversity of attributes is informed by our desire to build a comprehensive two-dimensional base map of n-dimensional geographic space. The paper discusses how standard GIS methods and neural network processing are combined towards the creation of an alternative map of the United States. [Copyright &y& Elsevier]

Abstract: This paper is aimed at creating an empirical model for assessing failure potential of highway slopes, with a special attention to the failure characteristics of the highway slopes in the Alishan, Taiwan area prior to, and post, the 1999 Chi-Chi, Taiwan earthquake. The basis of the study is a large database of 955 slope records from four highways in the Alishan area. Artificial neural network (ANN) is utilized to “learn” from this database. The developed ANN model is then used to study the effect of the Chi-Chi earthquake on the slope failure characteristics in the Alishan area. Significant changes in the degrees of influence of several factors (variables) are found and possible reasons for such changes are discussed. The novelty of this paper lies in the fact that the developed ANN models are used as a tool to investigate the slope failure characteristics before and after the Chi-Chi earthquake. [Copyright &y& Elsevier]

Several computational intelligence components, namely neural networks (NNs), fuzzy sets, and genetic algorithms (GAs), have been applied separately or in combination to the process of remotely sensed data classification. By applying computational intelligence, we expect increased accuracy through the use of NNs, optimal NN structure and parameter determination via GAs, and transparency using fuzzy sets is expected. This paper systematically reviews and compares several configurations in the particular context of remote sensing for land cover. In addition, some of the configurations used here, such as NEFCASS and CANFIS, have few previous applications in the field. A comparison of the configurations is achieved by testing the different methods with exactly the same case-study data. A thorough assessment of results is performed by constructing an accuracy matrix for each training and testing data set. The evaluation of different methods is not only based on accuracy but also on compactness, completeness, and consistency. The architecture, produced rule set, and training parameters for the specific classification task are presented. Some comments and directions for future work are given. [ABSTRACT FROM AUTHOR]

Assessing TBM performance is an important parameter for the successful accomplishment of a tunnelling project. This paper presents an attempt to model the advance rate of tunnelling with respect to the geological and geotechnical site conditions. The model developed for this particular task is implemented through the use of an artificial neural network (ANN) that allows the identification and understanding of both the way and the extent that the involved parameters affect the tunnelling process. The model described in the paper is customised for the construction of an interstation section of the Athens metro tunnels, where the ANN generalisations provided precise estimations regarding the anticipated advance rate. [Copyright &y& Elsevier]

In this paper, we propose the use of support vector machines (SVMs) for automatic hyperspectral data classification and knowledge discovery. In the first stage of the study, we use SVMs for crop classification and analyze their performance in terms of efficiency and robustness, as compared to extensively used neural and fuzzy methods. Efficiency is assessed by evaluating accuracy and statistical differences in several scenes. Robustness is analyzed in terms of: 1) suitability to working conditions when a feature selection stage is not possible and 2) performance when different levels of Gaussian noise are introduced at their inputs. In the second stage of this work, we analyze the distribution of the support vectors (SVs) and perform sensitivity analysis on the best classifier in order to analyze the significance of the input spectral bands. For classification purposes, six hyperspectral images acquired with the 128-band HyMAP spectrometer during the DAISEX-1999 campaign are used. Six crop classes were labeled for each image. A reduced set of labeled samples is used to train the models, and the entire images are used to assess their performance. Several conclusions are drawn: 1) SVMs yield better outcomes than neural networks regarding accuracy, simplicity, and robustness; 2) training neural and neurofuzzy models is unfeasible when working with high-dimensional input spaces and great amounts of training data; 3) SVMs perform similarly for different training subsets with varying input dimension, which indicates that noisy bands are successfully detected; and 4) a valuable ranking of bands through sensitivity analysis is achieved. [ABSTRACT FROM AUTHOR]

Geotechnical engineering deals with materials (e.g. soil and rock) that, by their very nature, exhibit varied and uncertain behavior due to the imprecise physical processes associated with the formation of these materials. Modeling the behavior of such materials in geotechnical engineering applications is complex and sometimes beyond the ability of most traditional forms of physically-based engineering methods. Artificial intelligence (AI) is becoming more popular and particularly amenable to modeling the complex behavior of most geotechnical engineering applications because it has demonstrated superior predictive ability compared to traditional methods. This paper provides state-of-the-art review of some selected AI techniques and their applications in pile foundations, and presents the salient features associated with the modeling development of these AI techniques. The paper also discusses the strength and limitations of the selected AI techniques compared to other available modeling approaches.

Copyright of Journal of Engineering & Architectural Faculty of Eskisehir Osmangazi University / Eskişehir Osmangazi Üniversitesi Mühendislik ve Mimarlık Fakültesi Dergisi is the property of Eskisehir Osmangazi University and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

In the current scenario, worldwide renewable energy systems receive renewed interest because of the global reduction of greenhouse gas emissions. This paper proposes a long-term wind speed prediction model based on various artificial neural network approaches such as Improved Back-Propagation Network (IBPN), Multilayer Perceptron Network (MLPN), Recursive Radial Basis Function Network (RRBFN), and Elman Network with five inputs such as wind direction, temperature, relative humidity, precipitation of water content and wind speed. The proposed ANN-based wind speed forecasting models help plan, integrate, and control power systems and wind farms. The simulation result confirms that the proposed Recursive Radial Basis Function Network (RRBFN) model improves the wind speed prediction accuracy and minimizes the error to a minimum compared to other proposed IBPN, MLPN, and Elman Network-based wind speed prediction models.

Although the principal aim of the rockfall management is to prevent rock boulders from reaching the buildings instead of the buildings resisting the boulder impacts, there usually exists a residual risk that has to be assessed, even when structural protection measurements are taken. The evaluation of the expected damage of buildings due to rockfalls using empirical data from past events is not always possible, as transferring and applying damage observations from one area to another can be unrealistic. In order to simulate potential rockfall scenarios and their damage on buildings, numerical methods can be an alternative. However due to their increased requirements in expertise and computational costs, their integration into the risk analysis is limited, and simpler tools to assess the rockfall vulnerability of buildings are needed. This paper focuses on the application of artificial intelligence AI methods for providing the expected damage of masonry walls which are subjected to rockfall impacts. First, a damage database with 672 datasets was created numerically using the particle finite element method and the finite element method. The input variables are the rock volume (VR), the rock velocity (RV), the masonry wall (t) and the masonry tensile strength fm. The output variable is a damage index (DI) equal to the percentage of the damaged wall area. Different AI algorithms were investigated and the ANN LM 4-21-1 model was selected to optimally assess the expected wall damage. The optimum model is provided here (a) as an analytical equation and (b) in the form of contour graphs, mapping the DI value. Known the VR and the RV, the DI can be directly used as an input for the vulnerability of masonry walls into the quantitative rockfall risk assessment equation.

Insights from a geoscience communication activity, verified using preliminary investigations with an artificial neural network, illustrate that observation of humans' abilities can help design an effective artificial intelligence or "AI". Even given only one set of "training" examples, survey participants could visually recognize which flow conditions created bedforms (e.g. sand dunes and riverbed ripples) from their shapes, but an interpreter's geoscience expertise does not help. Together, these observations were interpreted as indicating that a machine learning algorithm might be trained successfully from limited data, particularly if it is "helped" by pre-processing bedforms into a simple shape familiar from childhood play. [ABSTRACT FROM AUTHOR]

Despite the increasingly successful application of neural networks to many problems in the geosciences, their complex and nonlinear structure makes the interpretation of their predictions difficult, which limits model trust and does not allow scientists to gain physical insights about the problem at hand. Many different methods have been introduced in the emerging field of eXplainable Artificial Intelligence (XAI), which aims at attributing the network's prediction to specific features in the input domain. XAI methods are usually assessed by using benchmark datasets (such as MNISTor ImageNet for image classification). However, an objective, theoretically derived ground truth for the attribution is lacking for most of these datasets, making the assessment of XAI in many cases subjective. Also, benchmark datasets specifically designed for problems in geosciences are rare. Here, we provide a framework, based on the use of additively separable functions, to generate attribution benchmark datasets for regression problems for which the ground truth of the attribution is known a priori. We generate a large benchmark dataset and train a fully connected network to learn the underlying function that was used for simulation. We then compare estimated heatmaps from different XAI methods to the ground truth in order to identify examples where specific XAI methods perform well or poorly. We believe that attribution benchmarks as the ones introduced herein are of great importance for further application of neural networks in the geosciences, and for more objective assessment and accurate implementation of XAI methods, which will increase model trust and assist in discovering new science. [ABSTRACT FROM AUTHOR]

Abstract: Uncertain ground conditions represent the primary source of risk in underground tunnel construction. However, this problem can be solved by developing an accurate, probabilistic description of the geology. This paper presents a general model for probability based determination of tunnel geology that can be used as a basis for developing more effective decision support systems for tunneling design and construction. The proposed model is based on a Hidden Markov Model (HMM) and a neural network (NN). An approximate inference technique – a Particle Filter (PF) Algorithm – is used to simulate the geological parameters. This model overcomes the deficiencies of existing models by readily incorporating all available geologic information and updating geologic predictions based on observations given by the neural network. In order to validate the proposed model, the “Drainage Water Tunnel Project” at Zhong-He, Taipei, Taiwan was used. The results showed that the Neural-HMM model provides high accuracy in geological prediction. [Copyright &y& Elsevier]

Abstract: Developing a geological model is the first and a very important step during the reservoir simulation and modeling process. The geological model usually represents our best interpretation of the reservoir characteristics that extends beyond the well where we have actual measurements (logs, core, well test, etc.). The only real measurement with a large areal extent that geoscientists have access to is seismic data. Therefore, using seismic data to populate the high-resolution geological model is becoming increasingly popular. Using reservoir characteristics at the wellbore as the control point helps geoscientists in measuring the goodness of the correlation they create between seismic data and well logs. This paper presents a unique approach in accomplishing this task. The uniqueness of this approach is based on the fact that (a) it reduces the complexity of the model building process by dividing a very complex problem into two slightly less complex problems (surface seismic to VSP and VSP to log—i.e., divide and conquer) and (b) it effectively employs a synthetic set of formations representing the actual sequence of geological layers in the field in order to build a model and learn from it, and then, apply the lessons learned to the model building process for the actual reservoir. The results show that this strategy proves to be successful. [ABSTRACT FROM AUTHOR]

In this paper a new approach to quantitative Laser-Induced Breakdown Spectroscopy (LIBS) analysis of silicate rocks is presented. The method is adapted from the Franzini and Leoni algorithm, a method widely used in X-Ray Fluorescence analysis for correcting the matrix effects in the determination of the composition of geological materials. To illustrate the features of the new method proposed, nine elements were quantified in 19 geological standards by building linear univariate calibration curves, linear multivariate calibration surfaces (PLS) and using Artificial Neural Networks. The results were then compared with the predictions derived from the application of the algorithm here proposed. It was found that the Franzini and Leoni approach gives results much more precise than linear uni- and multivariate approaches, and comparable with the ones derived from the application of Artificial Neural Networks. A definite advantage of the proposed approach is the possibility of building multivariate non-linear calibration surfaces using linear optimization algorithms, a feature which makes the application of the Franzini and Leoni method in LIBS analysis much simpler (and controllable) with respect to the algorithms based on Artificial Neural Networks. Unlabelled Image • The Franzini-Leoni method, used in XRF analysis, is adapted to LIBS. • 19 geological samples are analyzed with this method. • The method is compared with linear/non-linear univariate and multivariate approaches. • The non-linear model can be obtained using linear optimization algorithms. [ABSTRACT FROM AUTHOR]

Lightning geolocation is useful in a variety of applications, ranging from weather nowcasting to a better understanding of thunderstorm evolution processes and remote sensing of the ionosphere. Lightning-generated radio signals can be used in range estimation of lightning return strokes, for which the most commonly employed technique is the time difference of arrival in lightning detection networks. Though these instrument networks provide the most reliability and best accuracy, users without access to them can instead benefit from lightning geolocation using a standalone instrument. In this article, we present the framework for training fast models capable of estimating negative cloud-to-ground lightning location from single-instrument observations of very low frequency/low frequency (VLF/LF, 3–300 kHz) radio pulses or “sferics,” without knowledge of the ionosphere’s D-region state. The models are generated using an analytical method, based on the delay between ground and skywave, and a machine learning method. The training framework is applied to three different data sets to assess model accuracy. Validation of the machine-learned models for these data sets, which include both simulated and observed sferics, confirms this technique as a promising solution for lightning distance estimation using a single receiver. Distance estimates using a machine-learned model for observed sferics in Kansas yield an RMSE of 53 km with 68% of them being within 9.8 km. Estimates using the analytical method are found to have an RMSE of 54 km with 68% of them being within 32 km. Limitations of our methodology and potential improvements to be investigated are also discussed. [ABSTRACT FROM AUTHOR]

Nowadays, artificial neural networks have been successfully applied on several research and application fields. An appropriate configuration for a neural network is a complex task, and it often requires the knowledge of an expert on the application. A technique for automatic configuration for a neural network is formulated as an optimization problem. Two strategies are considered: a mono-objective minimization problem, using multi-particle collision algorithm (MPCA); and a multi-objective minimization problem addressed by the non-dominated sorting genetic algorithm (NSGA-II). The proposed optimization approaches were tested for two application in geosciences: data assimilation for wave evolution equation, and the mesoscale seasonal climate prediction for precipitation. Better results with automatic configuration were obtained for data assimilation than those obtained by network defined by an expert. For climate seasonal precipitation, automatic configuration presented better predictions were presented than ones carried out by an expert. For the worked examples, the NSGA-II presented a superior result for the worked experiments. [ABSTRACT FROM AUTHOR]

Simulations and seismic inversions exhibit good performance in reservoir modeling task for the steady performance of conventional techniques. However, they still hardly meet the high demand of petroleum exploration since the impossibility of reaching high resolution both in vertical and lateral directions. Furthermore, simulations can only provide high-resolution results near loggings, while seismic inversions usually contain band-limited problems. Therefore, we present the modified generative adversarial nets with a decoder (DeGAN) as a novel approach, which is integrated with sequential simulation to realize super-resolution reservoir simulation. Specifically, the proposed method provides a geological model with high vertical resolution and optimized by the Zeoppritz function, introducing logging and seismic data simultaneously. After resampling and warping, DeGAN can be trained by these data sets and supplies a structure to generate high-frequency parts for reconstructing a super-resolution simulation of a subsurface profile. The proposed method presents the three-flow architecture of DeGAN to balance the contributions of three neural network models and utilizes this strategy in an offshore area successfully. By introducing multiple data sets, density experiments demonstrate that this approach can provide density profile with super-resolution for revealing possible thin layers, and the frequency distribution is in accord with loggings. The positive result verifies the effectiveness of this approach for providing a super-resolution simulation to supply a solution to the problem of the band-limited profile in seismic inversion. [ABSTRACT FROM AUTHOR]

Predicting the range of achievable strength and stiffness from stabilized soil mixtures is critical for engineering design and construction, especially for organic soils, which are often considered “unsuitable” due to their high compressibility and the lack of knowledge about their mechanical behavior after stabilization. This study investigates the mechanical behavior of stabilized organic soils using machine learning (ML) methods. ML algorithms were developed and trained using a database from a comprehensive experimental study (see Part I), including more than one thousand unconfined compression tests on organic clay samples stabilized by wet soil mixing (WSM) technique. Three different ML methods were adopted and compared, including two artificial neural networks (ANN) and a linear regression method. ANN models proved reliable in the prediction of the stiffness and strength of stabilized organic soils, significantly outperforming linear regression models. Binder type, mixing ratio, soil organic and water content, sample size, aging, temperature, relative humidity, and carbonation were the control variables (input parameters) incorporated into the ML models. The impacts of these factors were evaluated through rigorous ANN-based parametric analyses. Additionally, the nonlinear relations of stiffness and strength with these parameters were developed, and their optimum ranges were identified through the ANN models. Overall, the robust ML approach presented in this paper can significantly improve the mixture design for organic soil stabilization and minimize the experimental cost for implementing WSM in engineering projects.

Crab species are economically and ecologically important in coastal ecosystems, and their spatial distributions are pivotal for conservation and fisheries management. This study was focused on modelling the spatial distributions of three Portunidae crabs (Charybdis bimaculata, Charybdis japonica, and Portunus trituberculatus) in Haizhou Bay, China. We applied three analytical approaches (Generalized additive model (GAM), random forest (RF), and artificial neural network (ANN)) to spring and fall bottom trawl survey data (2011, 2013–2016) to develop and compare species distribution models (SDMs). Model predictability was evaluated using cross-validation based on the observed species distribution. Results showed that sea bottom temperature (SBT), sea bottom salinity (SBS), and sediment type were the most important factors affecting crab distributions. The relative importance of candidate variables was not consistent among species, season, or model. In general, we found ANNs to have less stability than both RFs and GAMs. GAMs overall yielded the least complex response curve structure. C. japonica was more pronounced in southwestern portion of Haizhou Bay, and C. bimaculata tended to stay in offshore areas. P. trituberculatus was the least region-specific and exhibited substantial annual variations in abundance. The comparison of multiple SDMs was informative to understand species responses to environmental factors and predict species distributions. This study contributes to better understanding the environmental niches of crabs and demonstrates best practices for the application of SDMs for management and conservation planning. [ABSTRACT FROM AUTHOR]

Abstract  A case study for the use of an artificial neural network (ANN) model for landslide susceptibility mapping in Koyulhisar (Sivas-Turkey) is presented. Digital elevation model (DEM) was first constructed using ArcGIS software. Relevant parameter maps were created, including geology, faults, drainage system, topographical elevation, slope angle, slope aspect, topographic wetness index, stream power index, normalized difference vegetation index and distance from roads. Finally, a landslide susceptibility map was constructed using the neural networks. The drawbacks of the method are discussed but as the validation procedures used confirmed the quality of the map produced, it is recommended the use of ANN may be helpful for planners and engineers in the initial assessment of landslide susceptibility. [ABSTRACT FROM AUTHOR]

Robust variants of nonlinear canonical correlation analysis (NLCCA) are introduced to improve performance on datasets with low signal-to-noise ratios, for example those encountered when making seasonal climate forecasts. The neural network model architecture of standard NLCCA is kept intact, but the cost functions used to set the model parameters are replaced with more robust variants. The Pearson product-moment correlation in the double-barreled network is replaced by the biweight midcorrelation, and the mean squared error (mse) in the inverse mapping networks can be replaced by the mean absolute error (mae). Robust variants of NLCCA are demonstrated on a synthetic dataset and are used to forecast sea surface temperatures in the tropical Pacific Ocean based on the sea level pressure field. Results suggest that adoption of the biweight midcorrelation can lead to improved performance, especially when a strong, common event exists in both predictor/predict and datasets. Replacing the mse by the mae leads to improved performance on the synthetic dataset, but not on the climate dataset except at the longest lead time, which suggests that the appropriate cost function for the inverse mapping networks is more problem dependent. [ABSTRACT FROM AUTHOR]