Search

Keywords provide the reader with a summary of the contents of the document and play a significant role in information retrieval systems, especially in search engine optimization and bibliographic databases. Furthermore keywords help to classify the document into the related topic. Keywords extraction included manual extracting depends on the content of the document or article and the judgment of its author. Manual extracting of keywords is costly, consumes effort and time, and error probability. In this research an automatic Arabic keywords extraction model based on deep learning algorithms is proposed. The model consists of three main steps: preprocessing, feature extraction and classification to classify the document tokens into keyword or not, Conventional Neural Networks (CNN) is used as a classifier. Two types of dataset are building in this research to test the proposed model, the first dataset form Arab Journal for Scientific Publishing (AJSP), the other dataset from Jordan Journal of Social Sciences (JJSS). The experiment results indicate promising results in the field of Arabic keyword extraction; the average accuracy of Conventional Neural Networks is found 0.97 with average precision 0.92. [ABSTRACT FROM AUTHOR]

A recent paper demonstrated that the pattern of firing rates across ~100 neurons in the anterior medial face patch is closely related to which human face (of 2,000) had been presented to a monkey [Chang L, Tsao DY (2017) Cell 169:1013-1028]. In addition, the firing rates for these neurons can be predicted for a novel human face. Although it is clear from this work that the firing rates of these face patch neurons encode faces, the properties of the face code have not yet been fully described. Based on an analysis of 98 neurons responding to 2,000 faces, I conclude that the anterior medial face patch uses a combinatorial rate code, one with an exponential distribution of neuron rates that has a mean rate conserved across faces. Thus, the face code is maximally informative (technically, maximum entropy) and is very similar to the code used by the fruit fly olfactory system. [ABSTRACT FROM AUTHOR]

Several high-dimensional models of Hopfield neural networks, such as complex-valued and quaternionic Hopfield neural networks, have been proposed. However, it has been hard to construct three-dimensional models of Hopfield neural networks. A split type of vector product Hopfield neural network (VPHNN) was proposed as a special case of ordinary Hopfield neural networks. It is easier to construct split types of Hopfield neural networks than multistate types of ones, because the split types can be often regarded as special cases of ordinary ones. In the present work, we extend the split VPHNN to the multistate VPHNN. We define its energy and a primitive learning algorithm, the Hebbian learning rule. In addition, we prove the stability of multistate VPHNNs. Furthermore, we investigated the fundamentals of multistate VPHNNs, such as the storage capacity and noise tolerance, by computer simulations. [ABSTRACT FROM AUTHOR]

In this letter, we propose a protocol for an automatic food recognition system that identifies the contents of the meal from the images of the food. We developed a multilayered convolutional neural network (CNN) pipeline that takes advantages of the features from other deep networks and improves the efficiency. Numerous traditional handcrafted features and methods are explored, among which CNNs are chosen as the best performing features. Networks are trained and fine-tuned using preprocessed images and the filter outputs are fused to achieve higher accuracy. Experimental results on the largest real-world food recognition database ETH Food-101 and newly contributed Indian food image database demonstrate the effectiveness of the proposed methodology as compared to many other benchmark deep learned CNN frameworks. [ABSTRACT FROM AUTHOR]

Copyright of Transformations in Business & Economics is the property of Vilnius 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.)

Neural networks, as powerful tools for data mining and knowledge engineering, can learn from data to build feature-based classifiers and nonlinear predictive models. Training neural networks involves the optimization of nonconvex objective functions, and usually, the learning process is costly and infeasible for applications associated with data streams. A possible, albeit counterintuitive, alternative is to randomly assign a subset of the networks' weights so that the resulting optimization task can be formulated as a linear least-squares problem. This methodology can be applied to both feedforward and recurrent networks, and similar techniques can be used to approximate kernel functions. Many experimental results indicate that such randomized models can reach sound performance compared to fully adaptable ones, with a number of favorable benefits, including (1) simplicity of implementation, (2) faster learning with less intervention from human beings, and (3) possibility of leveraging overall linear regression and classification algorithms (e.g., ℓ1 norm minimization for obtaining sparse formulations). This class of neural networks attractive and valuable to the data mining community, particularly for handling large scale data mining in real-time. However, the literature in the field is extremely vast and fragmented, with many results being reintroduced multiple times under different names. This overview aims to provide a self-contained, uniform introduction to the different ways in which randomization can be applied to the design of neural networks and kernel functions. A clear exposition of the basic framework underlying all these approaches helps to clarify innovative lines of research, open problems, and most importantly, foster the exchanges of well-known results throughout different communities. WIREs Data Mining Knowl Discov 2017, 7:e1200. doi: 10.1002/widm.1200 [ABSTRACT FROM AUTHOR]

The results of the study of the individual characteristics of pilots control actions in landing mode are presented. The study was carried out with the use of MiG-AT aircraft simulation facility. It is shown that the pilot inherent individual style of controlling aircraft in landing mode can be described with the individually adapted neural network model. This model for any moment of performing landing mode is specific to the pilot and presents correlation between the current state vector of “pilot-aircraft” system and predicted aircraft deviation from the Ideal landing coordinates. The simulation of landing trajectories on MiG-AT aircraft simulation facility was performed with the assistance of a group of operators who had different experience and qualifications. The results of these studies suggest that a particular characteristic of the pilot to control the aircraft in landing mode is uniquely manifested in the architecture of the neural network, the values of its parameters and the composition of the input signals. [ABSTRACT FROM AUTHOR]

While the adult human brain has approximately 8.8 × 1010 neurons, this number is dwarfed by its 1 × 1015 synapses. From the point of view of neuromorphic engineering and neural simulation in general this makes the simulation of these synapses a particularly complex problem. SpiNNaker is a digital, neuromorphic architecture designed for simulating large-scale spiking neural networks at speeds close to biological real-time. Current solutions for simulating spiking neural networks on SpiNNaker are heavily inspired by work on distributed high-performance computing. However, while SpiNNaker shares many characteristics with such distributed systems, its component nodes have much more limited resources and, as the system lacks global synchronization, the computation performed on each node must complete within a fixed time step. We first analyze the performance of the current SpiNNaker neural simulation software and identify several problems that occur when it is used to simulate networks of the type often used to model the cortex which contain large numbers of sparsely connected synapses. We then present a new, more flexible approach for mapping the simulation of such networks to SpiNNaker which solves many of these problems. Finally we analyze the performance of our new approach using both benchmarks, designed to represent cortical connectivity, and larger, functional cortical models. In a benchmark network where neurons receive input from 8000 STDP synapses, our new approach allows 4× more neurons to be simulated on each SpiNNaker core than has been previously possible. We also demonstrate that the largest plastic neural network previously simulated on neuromorphic hardware can be run in real time using our new approach: double the speed that was previously achieved. Additionally this network contains two types of plastic synapse which previously had to be trained separately but, using our new approach, can be trained simultaneously. [ABSTRACT FROM AUTHOR]

In recent years, the investors pay major attention to invest in gold market because of huge profits in the future. Gold is the only commodity which maintains its value even in the economic and financial crisis. Also, the gold prices are closely related with other commodities. The future gold price prediction becomes the warning system for the investors due to unforeseen risk in the market. Hence, an accurate gold price forecasting is required to foresee the business trends. This paper concentrates on forecasting the future gold prices from four commodities like historical data's of gold prices, silver prices, Crude oil prices, Standard and Poor's 500 stock index (S&P500) index and foreign exchange rate. The period used for the study is from 1st January 2000 to 31st April 2014. In this paper, a learning algorithm for single hidden layered Feed forward neural networks called Extreme Learning Machine (ELM) is used which has good learning ability. Also, this study compares the five models namely Feed forward networks without feedback, Feed forward back propagation networks, Radial basis function, ELMAN networks and ELM learning model. The results prove that the ELM learning performs better than the other methods. [ABSTRACT FROM AUTHOR]

A methodological framework for multiple steps ahead parking availability prediction is presented. Two different types of predictions are provided: the probability of a free space to continue being free in subsequent time intervals, and the short-term parking occupancy prediction in selected regions of an urban road network. The available data come from a wide network of on-street parking sensors in the “smart” city of Santander, Spain. The sensor network is segmented in four different regions, and then survival and neural network models are developed for each region separately. Findings show that the Weibull parametric models best describe the probability of a parking space to continue to be free in the forthcoming time intervals. Moreover, simple genetically optimized multilayer perceptrons accurately predict region parking occupancy rates up to 30 minutes in the future by exploiting 1-minute data. Finally, the real time, Web-based, implementation of the proposed parking prediction availability system is presented. [ABSTRACT FROM AUTHOR]

Brain ageing is followed by changes of the connectivity of white matter (WM) and changes of the grey matter (GM) concentration. Neurodegenerative disease is more vulnerable to an accelerated brain ageing, which is associated with prospective cognitive decline and disease severity. Accurate detection of accelerated ageing based on brain network analysis has a great potential for early interventions designed to hinder atypical brain changes. To capture the brain ageing, we proposed a novel computational approach for modeling the 112 normal older subjects (aged 50–79 years) brain age by connectivity analyses of networks of the brain. Our proposed method applied principal component analysis (PCA) to reduce the redundancy in network topological parameters. Back propagation artificial neural network (BPANN) improved by hybrid genetic algorithm (GA) and Levenberg–Marquardt (LM) algorithm is established to model the relation among principal components (PCs) and brain age. The predicted brain age is strongly correlated with chronological age ( r = 0.8). The model has mean absolute error (MAE) of 4.29 years. Therefore, we believe the method can provide a possible way to quantitatively describe the typical and atypical network organization of human brain and serve as a biomarker for presymptomatic detection of neurodegenerative diseases in the future. [ABSTRACT FROM AUTHOR]

A probabilistic neural-network-based feature-matching algorithm for a stereo image pair is presented in this paper, which will be useful as a constraint initializing method for further dense matching technique. In this approach, scale-invariant feature transform (SIFT) features are used to detect interest points in a stereo image pair. The descriptor which is associated with each keypoint is based on the histogram of the gradient magnitude and direction of gradients. These descriptors are the preliminary input for the matching algorithm. Using disparity range computed by visual inspection, the search area can be restricted for a given stereo image pair. Reduced search area improves the computation speed. Initial probabilities of matches are assigned to the keypoints which are considered as probable matches from the selected search area by Bayesian reasoning. The probabilities of all such matches are improved iteratively using relaxation labeling technique. Neighboring probable matches are exploited to improve the probability of best match using consistency measures. Confidence measures considering the neighborhood, unicity, and symmetry are some validation techniques which are built into the technique presented here for finding accurate matches. The algorithm is found to be effective in matching SIFT features detected in a stereo image pair with greater accuracy, and these accurate correspondences can be used in finding the fundamental matrix which encodes the epipolar geometry between the given stereo image pair. This fundamental matrix can then be used as a constraint for finding inliers that are used in matching methods for deriving dense disparity map. [ABSTRACT FROM PUBLISHER]

Copyright of Iraqi Journal of Statistical Sciences is the property of Republic of Iraq Ministry of Higher Education & Scientific Research (MOHESR) 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.)

This paper deals with a new transformation and fusion of digital input patterns used to train and test feedforward neural network for a wound-rotor three-phase induction machine windings short-circuit diagnosis. The single type of short-circuits tested by the proposed approach is based on turn-to-turn fault which is known as the first stage of insulation degradation. Used input/output data have been binary coded in order to reduce the computation complexity. A new procedure, namely addition and mean of the set of same rank, has been implemented to eliminate the redundancy due to the periodic character of input signals. However, this approach has a great impact on the statistical properties on the processed data in terms of richness and of statistical distribution. The proposed neural network has been trained and tested with experimental signals coming from six current sensors implemented around a setup with a prime mover and a 5.5 kW wound-rotor three-phase induction generator. Both stator and rotor windings have been modified in order to sort out first and last turns in each phase. The experimental results highlight the superiority of using this new procedure in both training and testing modes. [ABSTRACT FROM PUBLISHER]

The procedure for the synthesis method of neural network prediction of external perturbations on the special-purpose entity (SPE) is presented. A comparison of the time series forecasting ambient temperature using a typical approach (gradient method) to optimize the weighting coefficients of the neural network and genetic algorithm. Investigated the quality of the system control unit neural network prediction and optimization of its configuration based on genetic algorithm compared to a system without such a prediction. Found that the solution to the problem of mathematical modeling of time series of projections can significantly improve the performance of the control system and to maximize the efficiency of the SPE. [ABSTRACT FROM AUTHOR]

Currently, remote sensing technologies were widely employed in the dynamic monitoring of the land. This paper presented an algorithm named fuzzy nonlinear proximal support vector machine (FNPSVM) by basing on ETM+ remote sensing image. This algorithm is applied to extract various types of lands of the city Da’an in northern China. Two multi-category strategies, namely “one-against-one” and “one-against-rest” for this algorithm were described in detail and then compared. A fuzzy membership function was presented to reduce the effects of noises or outliers on the data samples. The approaches of feature extraction, feature selection, and several key parameter settings were also given. Numerous experiments were carried out to evaluate its performances including various accuracies (overall accuracies and kappa coefficient), stability, training speed, and classification speed. The FNPSVM classifier was compared to the other three classifiers including the maximum likelihood classifier (MLC), back propagation neural network (BPN), and the proximal support vector machine (PSVM) under different training conditions. The impacts of the selection of training samples, testing samples and features on the four classifiers were also evaluated in these experiments. [ABSTRACT FROM AUTHOR]

This paper presents a novel algorithm based on generalized opposition-based learning (GOBL) to improve the performance of differential evolution (DE) to solve high-dimensional optimization problems efficiently. The proposed approach, namely GODE, employs similar schemes of opposition-based DE (ODE) for opposition-based population initialization and generation jumping with GOBL. Experiments are conducted to verify the performance of GODE on 19 high-dimensional problems with D = 50, 100, 200, 500, 1,000. The results confirm that GODE outperforms classical DE, real-coded CHC (crossgenerational elitist selection, heterogeneous recombination, and cataclysmic mutation) and G-CMA-ES (restart covariant matrix evolutionary strategy) on the majority of test problems. [ABSTRACT FROM AUTHOR]

Copyright of Economía del Caribe is the property of Fundacion Universidad del Norte 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.)

Abstract: This paper presents a system for recognizing static gestures of alphabets in Persian sign language (PSL) using Wavelet transform and neural networks (NN). The required images for the selected alphabets are obtained using a digital camera. The color images are cropped, resized, and converted to grayscale images. Then, the discrete wavelet transform (DWT) is applied on the gray scale images, and some features are extracted. Finally, the extracted features are used to train a Multi-Layered Perceptron (MLP) NN. Our recognition system does not use any gloves or visual marking systems. This system only requires the images of the bare hand for the recognition. The system is implemented and tested using a data set of 640 samples of Persian sign images; 20 images for each sign. Experimental results show that our system is able to recognize 32 selected PSL alphabets with an average classification accuracy of 94.06%. [ABSTRACT FROM AUTHOR]

A great many studies on wave breaking have been carried out, and much experimental and field data have been documented. Moreover, on the basis of various data sets, many empirical formulas based primarily on regression analysis have been proposed to quantitatively estimate wave breaking for engineering applications. However, wave breaking has an inherent variability, which suggests that a linear statistical approach such as regression analysis might be inadequate. This study presents an alternative nonlinear method using an artificial neural network (ANN), one of the soft computing methods, for predicting breaking-wave heights and water depths. Using data from laboratory experiments showing that wave breaking characteristics on a gravel beach are different from those on a sandy beach, we developed a three-layered feed-forward type of network to obtain the output of wave-breaking heights and water depths using deepwater heights, wave periods, and seabed conditions as inputs. In particular, the effects of the groundwater table on wave breaking on a gravel beach were considered. We then aggregated the laboratory data to train the ANN to predict the wave breaker index. The results predicted by fully-trained ANNs are more accurate than those obtained by existing empirical formulas, and they identified factors crucial for predicting wave breaking. These results show that ANNs are useful tools for both the prediction and mechanistic analysis of wave breaking. [ABSTRACT FROM AUTHOR]

One information technology that may be considered by transportation managers, and which is included in the portfolio of technologies that encompass TMS, is artificial neural networks (ANNs). These artificially intelligent computer decision support software provide solutions by finding and recognizing complex patterns in data. ANNs have been used successfully by transportation managers to forecast transportation demand, estimate future transport costs, schedule vehicles and shipments, route vehicles and classify carriers for selection. Artificial neural networks excel in transportation decision environments that are dynamic, complex and unstructured. This article introduces ANNs to transport managers by describing ANN technological capabilities, reporting the current status of transportation neural network applications, presenting ANN applications that offer significant potential for future development and offering managerial guidelines for ANN development. [ABSTRACT FROM AUTHOR]

The article examines the roughness and curvature deviation of nonlinear fine optimal lens grinding parameters using the combined merits of the Taguchi method and neural network software. It notes that Taguchi method with an L18 orthogonal array was used in the first stage optimization while ANNODE software for the second optimization. Results confirm that lens surfaces ground with polishers ranging in curvature from -7.00 to +7.00 are sturdy and strong in desired quality targets.

The article presents a case study which aims to create a usable life forecast model for consumable parts in semiconductor manufacturing with the use of neural network. The study is focusing on a consumable probe card used in the semiconductor wafer testing operation. The study applied a neural network analysis software to forecast the usable life by putting the initial values of the neural network variables.

In this study, a microcontroller-based gas mixture classification system is proposed to use real-time analyses of the trichloroethylene and acetone binary mixture. A Feed Forward Neural Network (FFNN) structure is performed for quantitative identification of individual gas concentrations (trichloroethylene and acetone) in their gas mixtures. The phthalocyaninecoated Quartz Crystal Microbalance (QCM) type sensors were used as gas sensors. A calibrated Mass Flow Controller (MFC) was used to control the flow rates of carrier gas and trichloroethylene and acetone gas mixtures streams. The components in the binary mixture were quantified by applying the sensor responses from the QCMs sensor array as inputs to the FFNN. The microcontroller-based gas mixture classification system performs Neural Network (NN)-based estimation, the data acquisition and user interface tasks. This system can estimate the gas concentrations of trichloroethylene and acetone with the average errors of 0.08% and 0.97%, respectively. [ABSTRACT FROM AUTHOR]

Recently we developed a model for prediction of pH-dependent aqueous solubility of drugs and drug like molecules. In the present work, the model was applied on a series of novel Histone Deacetylases (HDAC) inhibitors discovered at TopoTarget. The applicability of our model was evaluated on the series of HDAC inhibitors by use of Self-Organizing Maps (SOM) and 2D-projection of the HDAC inhibitors on the chemical space of the training data set of the artificial neural network (ANN) module. The model was refined for the particular chemical space of interest, which led to two modifications in the training data set of the ANN. The performance of the original and the two modified versions of the model were evaluated against the commercial software from Simulations-plus and pH-dependent solubility measurements for representative compounds of the series. The results of the evaluation indicate that one can develop models that are more accurate in predicting differences in the solubility of structurally very similar compounds than models that have been trained on structurally unbiased, diverse data sets. Such 'tailor-made' models have the potential to become trustworthy enough to replace time-consuming and expensive medium- and high-throughput solubility experiments by providing results of similar or even better quality. [ABSTRACT FROM AUTHOR]

This paper describes the design, implementation and testing of the software tool designated as UV-SRNA-PDA (Universidad del Valle's artificial neural network simulator for a personal digital assistant), which is oriented to artificial neural networks simulation and complex industrial processes identification. This application works on a PDA (personal digital assistant) Palm T5 using a customized data acquisition system. Two kinds of artificial neural networks were implemented: perceptron and multi-layer perceptron (MLP) using as learning algorithms the following: backpropagation, descendent gradient, variable learning rate and momentum. To test the proposed system, first and second order systems were selected (only the latter is reported here), finding their neural models and validating their results with MATLAB and UV-SRNA 2.0 (PC version of UV- SRNA-PDA). These tests achieved an average training error of 5.62 x 10-3 ± 3.55 x 10-4 and an average validating error of 4.56 x 10-3 ± 5.95 x 10-4. In both cases, the results were better or comparable with those from other software simulation tools. However, the typical training time on the UV-SRNA-PDA was 900 s compared to 3 s in MATLAB and 8 s in UV-SRNA 2.0. [ABSTRACT FROM AUTHOR]

Robust and reliable vehicle detection from images acquired by a moving vehicle (i.e., on-road vehicle detection) is an important problem with applications to driver assistance systems and autonomous, self-guided vehicles. The focus of this work is on the issues of feature extraction and classification for rear-view vehicle detection. Specifically, by treating the problem of vehicle detection as a two-class classification problem, we have investigated several different feature extraction methods such as principal component analysis, wavelets, and Gabor filters. To evaluate the extracted features, we have experimented with two popular classifiers, neural networks and support vector machines (SVMs). Based on our evalnation results, we have developed an on-board real-time monocular vehicle detection system that is capable of acquiring grey-scale images, using Ford's proprietary low-light camera, achieving an average detection rate of 10 Hz. Our vehicle detection algorithm consists of two main steps: a multiscale driven hypothesis generation step and an appearance-based hypothesis verification step. During the hypothesis generation step, image locations where vehicles might be present are extracted. This step uses multiscale techniques not only to speed up detection, but also to improve system robustness. The appearance-based hypothesis verification step verifies the hypotheses using Gabor features and SVMs. The system has been tested in Ford's concept vehicle under different traffic conditions (e.g., structured highway, complex urban streets, and varying weather conditions), illustrating good performance. [ABSTRACT FROM AUTHOR]

A class of data-reusing learning algorithms for real-time recurrent neural networks (RNNs) is analyzed. The analysis is undertaken for a general sigmoid nonlinear activation function of a neuron for the real time recurrent learning training algorithm. Error bounds and convergence conditions for such data-reusing algorithms are provided for both contractive and expansive activation functions. The analysis is undertaken for various configurations that are generalizations of a linear structure infinite impulse response adaptive filter. [ABSTRACT FROM AUTHOR]

Presents an incremental learning algorithm for supervised neural networks. Accommodation of data; Access to used data during incremental learning sessions; Generation of multiple hypothesis using training data.

Empirical studies in inductive language learning point at pure memory-based learning as a successful approach to many language learning tasks, often performing better than learning methods that abstract from the learning material. The possibility is left open, however, that limited, careful abstraction in memory-based learning may be harmless to generalization, as long as the disjunctivity of language data is preserved. We test this hypothesis, by comparing empirically a range of careful abstraction methods, focusing particularly on methods that (i) generalize instances and (ii) perform oblivious (partial) decision-tree abstraction. These methods are applied to a selection of language learning tasks, and their generalization performance as well as memory item compression rates are collected. On the basis of the results we conclude that when combined with feature weighting or value distance metrics, careful abstraction equals or outperforms pure memory-based learning, yet mainly on small data sets. In the concluding case study involving large data sets, we find that the FAMBL algorithm, a new careful abstractor which merges families of instances, performs close to pure memory-based learning, though it equals it only on three of the six tasks. On the basis of the gathered empirical results, we discuss the incorporation of the notion of instance families, i.e. carefully generalized instances, in memory-based language learning. [ABSTRACT FROM AUTHOR]

Discusses teaching applications of several personal computer (PC) and workstation programs for simulating dynamic systems and neural methods. DESIRE computer software; FORTRAN; DESIRE/NEUNET; PC simulation under OS/2 and workstation programs for UNIX/OPEN Windows.

Reviews software for studying and developing applications of artificial neural networks. ExploreNet 3000; KnowledgeNet; DataBase Mining; NeuDesk 2; NeuModel; Neurun; AutoNet Windows; AutoNet CCode; AutoNet WorkBench; NeuralWorks Professional II/Plus; NeuralWorks Explorer Educational; 4Thought.

Detection of abnormalities in wireless capsule endoscopy (WCE) images is a challenging task. Typically, these images suffer from low contrast, complex background, variations in lesion shape and color, which affect the accuracy of their segmentation and subsequent classification. This research proposes an automated system for detection and classification of ulcers in WCE images, based on state-of-the-art deep learning networks. Deep learning techniques, and in particular, convolutional neural networks (CNNs), have recently become popular in the analysis and recognition of medical images. The medical image datasets used in this study were obtained from WCE video frames. In this work, two milestone CNN architectures, namely the AlexNet and the GoogLeNet are extensively evaluated in object classification into ulcer or non-ulcer. Furthermore, we examine and analyze the images identified as containing ulcer objects to evaluate the efficiency of the utilized CNNs. Extensive experiments show that CNNs deliver superior performance, surpassing traditional machine learning methods by large margins, which supports their effectiveness as automated diagnosis tools. [ABSTRACT FROM AUTHOR]

Artificial Intelligence (AI) has been widely used in Short-Term Load Forecasting (STLF) in the last 20 years and it has partly displaced older time-series and statistical methods to a second row. However, the STLF problem is very particular and specific to each case and, while there are many papers about AI applications, there is little research determining which features of an STLF system is better suited for a specific data set. In many occasions both classical and modern methods coexist, providing combined forecasts that outperform the individual ones. This paper presents a thorough empirical comparison between Neural Networks (NN) and Autoregressive (AR) models as forecasting engines. The objective of this paper is to determine the circumstances under which each model shows a better performance. It analyzes one of the models currently in use at the National Transport System Operator in Spain, Red Eléctrica de España (REE), which combines both techniques. The parameters that are tested are the availability of historical data, the treatment of exogenous variables, the training frequency and the configuration of the model. The performance of each model is measured as RMSE over a one-year period and analyzed under several factors like special days or extreme temperatures. The AR model has 0.13% lower error than the NN under ideal conditions. However, the NN model performs more accurately under certain stress situations. [ABSTRACT FROM AUTHOR]

The article reports that the Scientists have discovered a supercolony of Adélie penguins in the Danger Islands, Antarctica. Topics discussed include use of neural network software for analyzing penguin nests; adverse impact of climate change on population of Adélie penguins; and use of commercial quadcopter drone for taking image of Adélie penguins.

The article presents a study on the estimation of furrow irrigation sediment loss using an Artificial Neural Network (ANN) conducted by Bradley A. King and others. Topics discussed include elements of artificial neural network (ANN) modeling; training, testing, and validation of an ANN data; and challenges of an ANN model to discover the data generation process.

Rapid volatile profiling of stool sample headspace was achieved using a combination of short multi-capillary chromatography column (SMCC), highly sensitive heated metal oxide semiconductor sensor and artificial neural network software. For direct analysis of biological samples this prototype offers alternatives to conventional gas chromatography (GC) detectors and electronic nose technology. The performance was compared to an identical instrument incorporating a long single capillary column (LSCC). The ability of the prototypes to separate complex mixtures was assessed using gas standards and homogenized in house ‘standard’ stool samples, with both capable of detecting more than 24 peaks per sample. The elution time was considerably faster with the SMCC resulting in a run time of 10 min compared to 30 min for the LSCC. The diagnostic potential of the prototypes was assessed using 50 C. difficile positive and 50 negative samples. The prototypes demonstrated similar capability of discriminating between positive and negative samples with sensitivity and specificity of 85% and 80% respectively. C. difficile is an important cause of hospital acquired diarrhoea, with significant morbidity and mortality around the world. A device capable of rapidly diagnosing the disease at the point of care would reduce cases, deaths and financial burden. [ABSTRACT FROM AUTHOR]