Aneural Network Research Articles

Predicting the development of vulvar lichen sclerosus

Relevance. The issue of timely diagnosis and treatment of vulvar lichen sclerosus has become especially acute in recent years due to the “rejuvenation” of the disease and the risk of its malignancy. In this regard, it is urgent to search for effective methods for predicting and early detection of the disease. The aim of the study - to develop amodel for predicting vulvar lichen sclerosus based on established clinical and anamnestic risk factors. Materials and Methods. The prospective case-control study included 404 women aged 20 to 70 years, of which 344 were patients with vulvar lichen sclerosus and 60 were women without vulvar diseases. At the first stage, acomparative statistical correlation analysis of the clinical and anamnestic data of the subjects was carried out using the Spearman correlation coefficient (R0.15), Chi-square tests, Phi and Cramer statistics, the Mann-­Whitney U test and the Student t test (p 0.05). The data obtained were used to develop aneural network model for predicting vulvar lichen sclerosus in the second stage of the study. Results and Discussion. Based on established reliably significant (p0.05) obstetric-­gynecological, somatic, infectious, hygienic and household factors influencing the risk of developing vulvar lichen sclerosus (Rindicator - from 0.16 to 0.38 confirms the statistical significance of correlations), aneural network model for predicting vulvar lichen sclerosus was developed (the percentage of correct classification on the test sample is the maximum possible value - 100%) and acomputer program was written that automates the procedure for predicting the disease. Conclusion. The neural network model for predicting the disease, developed on the basis of reliably (p0.05) significant risk factors for vulvar lichen sclerosus, has high prognostic properties, and acomputer program written on its basis allows the doctor in amatter of minutes to identify the patient at risk for the development of vulvar lichen sclerosus and give she needs preventive recommendations aimed at preventing or early detection of the disease.

Computer Assisted Bone Age Estimation Using Dimensions of Metacarpal Bones and Metacarpophalangeal Joints Based on Neural Network.

Bone age is a more accurate assessment for biologic development than chronological age. The most common method for bone age estimation is using Pyle and Greulich Atlas. Today, computer-based techniques are becoming more favorable among investigators. However, the morphological features in Greulich and Pyle method are difficult to be converted into quantitative measures. During recent years, metacarpal bones and metacarpophalangeal joints dimensions were shown to be highly correlated with skeletal age. In this study, we have evaluated the accuracy and reliability of a trained neural network for bone age estimation with quantitative and recently introduced related data, including chronological age, height, trunk height, weight, metacarpal bones, and metacarpophalangeal joints dimensions. In this cross sectional retrospective study, aneural network, using MATLAB, was utilized to determine bone age by employing quantitative features for 304 subjects. To evaluate the accuracy of age estimation software, paired t-test, and inter-class correlation was used. The difference between the mean bone ages determined by the radiologists and the mean bone ages assessed by the age estimation software was not significant (p Value= 0.119 in male subjects and p= 0.922 in female subjects). The results from the software and radiologists showed a strong correlation -ICC=0.990 in male subjects and ICC=0.986 in female subjects (p< 0.001). The results have shown an acceptable accuracy in bone age estimation with training neural network and using dimensions of bones and joints.

Designing the architecture of a distributed system for information monitoring of IoT and IIoT infrastructures traffic

The article considers an approach to designing the architecture of a distributed information monitoring and quality management system for communication services provided by the Internet of Things and Industrial Internet of Things infrastructures based on solutions supporting machine-to-machine and human-machine interaction. The choice of a neural network with long Short-Term memory (LSTM) is justified as a theoretical basis for modeling the active node of an information monitoring system. The developed structure of the LSTM network is presented, taking into account the spatial and temporal correlation of the obtained indicators at various observation points of the distributed information monitoring system. The criteria for evaluating the effectiveness of the proposed model for the conditions of information monitoring of non-stationary load in a mode close to real time are substantiated.

Multi objective optimization for house roof using artificial neural network model

Roof models with low heat loss and low heating costs for buildings are crucial for reducing energy consumption and greenhouse gas emissions in cold regions. This study aimed to calculate the heat loss for various roof models with low heat loss, which are widely used in houses in Europe and other cold regions. We created 324 models for three types of roofs (light, medium, and heavy) with different materials and thicknesses using the HAP program. We then trained a neural network (ANN) to generate a mathematical function that can be used to calculate the optimal surface using a multi-objective genetic algorithm method. The resulting optimal roof design was simulated using the Ansys program on a January day. We compared the heat loss results for the optimal roof for HAP, ANN, and transient thermal analysis by Ansys, respectively, showing a close agreement among the methods.

BLDC Motors Sensorless Control Based on MLP Topology Neural Network

In order to reduce the complexity of the brushless DC motor (BLDC)-control-system algorithm while improving the estimation performance of the rotor phase position and the speed of the sensorless motor, a neural network (ANN) control algorithm based on multi-layer perceptron (MLP) topology is proposed. The phase voltage of the motor is conditioned to obtain the phase-voltage signal with a high signal-to-noise ratio, which is used as the input eigenvalue of the multi-layer-perceptron-topology neural network algorithm. The encoder signal is used as the training test data of the MLP-ANN. The first layer of the perceptual neural network estimates the position according to the voltage characteristics with incremental time characteristics. The second layer of the perceptual neural network estimates the speed according to the collected time characteristics and the characteristics of rotor position error. The algorithm after learning and training is digitally discretized and integrated into the motor control system. Experimental tests were carried out under no-load, speed step and load mutation conditions. The experimental results show that the algorithm can accurately estimate the rotor position and speed. The absolute error of the rotor position is within 0.02 rad, and the absolute error of the rotor speed is within 4 rpm. The control system with strong robustness has good dynamic and static characteristics.

Hardware Trojan Detection Using Machine Learning: A Tutorial

With the growth and globalization of IC design and development, there is an increase in the number of Designers and Design houses. As setting up a fabrication facility may easily cost upwards of $20 billion, costs for advanced nodes may be even greater. IC design houses that cannot produce their chips in-house have no option but to use external foundries that are often in other countries. Establishing trust with these external foundries can be a challenge, and these foundries are assumed to be untrusted. The use of these untrusted foundries in the global semiconductor supply chain has raised concerns about the security of the fabricated ICs targeted for sensitive applications. One of these security threats is the adversarial infestation of fabricated ICs with a Hardware Trojan (HT) . An HT can be broadly described as a malicious modification to a circuit to control, modify, disable, or monitor its logic. Conventional VLSI manufacturing tests and verification methods fail to detect HT due to the different and un-modeled nature of these malicious modifications. Current state-of-the-art HT detection methods utilize statistical analysis of various side-channel information collected from ICs, such as power analysis, power supply transient analysis, regional supply current analysis, temperature analysis, wireless transmission power analysis, and delay analysis. To detect HTs, most methods require a Trojan-free reference golden IC. A signature from these golden ICs is extracted and used to detect ICs with HTs. However, access to a golden IC is not always feasible. Thus, a mechanism for HT detection is sought that does not require the golden IC. Machine Learning (ML) approaches have emerged to be extremely useful in helping eliminate the need for a golden IC. Recent works on utilizing ML for HT detection have been shown to be promising in achieving this goal. Thus, in this tutorial, we will explain utilizing ML as a solution to the challenge of HT detection. Additionally, we will describe the Electronic Design Automation (EDA) tool flow for automating ML-assisted HT detection. Moreover, to further discuss the benefits of ML-assisted HT detection solutions, we will demonstrate a Neural Network (NN) -assisted timing profiling method for HT detection. Finally, we will discuss the shortcomings and open challenges of ML-assisted HT detection methods.

PV Panel Model Parameter Estimation by Using Neural Network

Photovoltaic (PV) panels have been widely used as one of the solutions for green energy sources. Performance monitoring, fault diagnosis, and Control of Operation at Maximum Power Point (MPP) of PV panels became one of the popular research topics in the past. Model parameters could reflect the health conditions of a PV panel, and model parameter estimation can be applied to PV panel fault diagnosis. In this paper, we will propose a new algorithm for PV panel model parameters estimation by using a Neural Network (ANN) with a Numerical Current Prediction (NCP) layer. Output voltage and current signals (VI) after load perturbation are observed. An ANN is trained to estimate the PV panel model parameters, which is then fined tuned by the NCP to improve the accuracy to about 6%. During the testing stage, VI signals are input into the proposed ANN-NCP system. PV panel model parameters can then be estimated by the proposed algorithms, and the estimated model parameters can be then used for fault detection, health monitoring, and tracking operating points for MPP conditions.

Exploiting Manifold Feature Representation for Efficient Classification of 3D Point Clouds

In this paper, we propose an efficient point cloud classification method via manifold learning based feature representation. Different from conventional methods, we use manifold learning algorithms to embed point cloud features for better considering the geometric continuity on the surface. Then, the nature of point cloud can be acquired in low dimensional space, and after being concatenated with features in the original three-dimensional (3D) space, both the capability of feature representation and the classification network performance can be improved. We explore three traditional manifold algorithms (i.e., Isomap, Locally-Linear Embedding, and Laplacian eigenmaps) in detail, and finally, we select the Locally-Linear Embedding (LLE) algorithm due to its low complexity and locality consistency preservation. Furthermore, we propose a neural network based manifold learning (NNML) method to implement manifold learning based non-linear projection. Experiments demonstrate that the proposed two manifold learning methods can obtain better performances than the state-of-the-art methods, and the obtained mean class accuracy (mA) and overall accuracy (oA) can reach 91.4% and 94.4%, respectively. Moreover, because of the improved feature learning capability, the proposed NNML method can also have better classification accuracy on models with prominent geometric shapes. To further demonstrate the advantages of PointManifold, we extend it as a plug and play method for point cloud classification task, which can be directly used with existing methods and gain a significant improvement.

Toward new tools for biodiversity studies: the use of portable near-infrared spectroscopy combined with machine learning to identify species of Decapoda

Context Accuracy in species identification is a crucial factor for the quality of biodiversity studies and species management. Ensuring high accuracy is challenging for diverse taxonomic groups, including those with fishery importance such as Decapoda. Aims The objective of the present study was to use portable near-infrared spectroscopy combined with machine learning through a neural network (ANN) to identify species of Decapoda. Methods We propose an ANN application that rapidly and accurately emulates the results that would be obtained by a specialist. We used 124 specimens from seven marine Decapoda species as a dataset to fit the model. Key results The ANN was able to correctly learn (classify) all the patterns of the species (100% accuracy), with an overall mean probability of 0.97 ± 0.068. Conclusions The results obtained using portable near-infrared spectroscopy combined with machine learning (ANN) demonstrated that this method can be used with high accuracy to distinguish Decapoda species. Implications Studies aiming at comparisons among species may consider the use of this technique for the precise and inexpensive separation among species by non-specialists or for species that require the identification of a large number of individuals.

Artificial Intelligence-Aided Low Cost and Flexible Graphene Oxide-Based Paper Sensor for Ultraviolet and Sunlight Monitoring

The adverse effect of ultraviolet (UV) radiation on human beings has sparked intense interest in the development of new sensors to effectively monitor UV and solar exposure. This paper describes a novel low-cost and flexible graphene oxide (GO)-based paper sensor capable of detecting the total amount of UV or sun energy delivered per unit area. GO is incorporated into the structure of standard printing paper, cellulose, via a low-cost fabrication technique. The effect of UV and solar radiation exposure on the GO paper-based sensor is investigated using a simple color change analysis. As a result, users can easily determine the amount of ultraviolet or solar energy received by the sensor using a simple color analysis application. A neural network (ANN) model is also explored to learn the relation between UV color intensity and exposure time, then digitally display the results. The accuracy for the developed ANN reached 96.83%. The disposable, cost-effective, simple, biodegradable, safe, and flexible characteristics of the paper-based UV sensor make it an attractive candidate for a variety of sensing applications. This work provides new vision toward developing highly efficient and fully disposable GO-based photosensors.Graphical

An approach to removing COD and BOD based on polycarbonate mixed matrix membranes that contain hydrous manganese oxide and silver nanoparticles: A novel application of artificial neural network based simulation in MATLAB

This study aimed to determine the efficacy of novel ultrafiltration and mixed matrix membrane (MMM) composed of hydrous manganese oxide (HMO) and silver nanoparticles (Ag-NPs) for the removal of biological oxygen demand (BOD) and chemical oxygen demand (COD). In the polycarbonate (PC) MMM, the weight percent of HMO and Ag-NP has been increased from 5% to 10%. A neural network (ANN) was used in this study to compare PC-HMO and Ag-NP. MMM was evaluated in combination with HMO and Ag-NP loadings in order to assess their effects on pure water flux, mean pore size, porosity, and efficacy in removing BOD and COD. HMO and Ag-NPs can decrease membrane porosity in the casting solution while increasing mean pore size. According to the study's findings, the artificial neural network model appears to be highly appropriate for predicting the removal of BOD and COD. To develop a successful model, a suitable input dataset was selected, which consisted of BOD and COD. An ideal model architecture for MMM was proposed based on an optimal number of hidden layers (2 layers) and neurons (5–8 neurons). Experiments and predicted data show a strong correlation between the developed models. BOD was predicted with an excellent R2 and a low root mean square error (RMSE) of 0.99 and 0.05%, respectively, while COD was predicted with an excellent R2 and a low RMSE of 0.99 and 0.09%, respectively. Based on the results, Ag-NP was found to be an excellent candidate for the preparation of MMMs as well as convenient for the removal of BOD and COD from polluted water sources.

A unified parameter model based on machine learning for describing microbial transport in porous media

The transport and retention of microorganisms are typically described using attachment/detachment and straining/liberation models. However, the parameters in the models varied significantly, posing a significant challenge to describe microbial transport under different environmental conditions. A neural network (ANN) model was developed in this study to link the parameters in the model with the factors influencing microbial transport including the properties of microorganisms such as size and surface potentials, and the properties of porous media such as grain size and porosity, and flow conditions. Exhaustive search of literature renders 420 sets of experimental data of microbial transport, which were fitted using the microbial transport model to obtain model parameters. The model parameters, together with the factors influencing microbial transport, were then used to train an ANN model to search for their relationship. An ANN-based parameter relationship was derived and was then used to simulate microbial transport. The simulated results using the relationship roughly matched with the experimental data under different environmental conditions, indicating that a unified relationship was established between the parameters of the microbial transport model and the factors influencing microbial transport, and that microbial transport can be described using the microbial transport model with the ANN-based unified relationship for model parameters.

Development and evaluation of adeep learning algorithm for German word recognition from lip movements

HintergrundZahlreiche Menschen profitieren beim Lippenlesen von den zusätzlichen visuellen Informationen aus den Lippenbewegungen des Sprechenden, was jedoch sehr fehleranfällig ist. Algorithmen zum Lippenlesen mit auf künstlichen neuronalen Netzwerken basierender künstlicher Intelligenz verbessern die Worterkennung signifikant, stehen jedoch nicht für die deutsche Sprache zur Verfügung.Material und MethodenEs wurden 1806 Videos mit jeweils nur einer deutsch sprechenden Person selektiert, in Wortsegmente unterteilt und mit einer Spracherkennungssoftware Wortklassen zugeordnet. In 38.391 Videosegmenten mit 32 Sprechenden wurden 18 mehrsilbige, visuell voneinander unterscheidbare Wörter zum Trainieren und Validieren eines neuronalen Netzwerks verwendet. Die Modelle 3D Convolutional Neural Network, Gated Recurrent Units und die Kombination beider Modelle (GRUConv) wurden ebenso verglichen wie unterschiedliche Bildausschnitte und Farbräume der Videos. Die Korrektklassifikationsrate wurde jeweils innerhalb von 5000 Trainingsepochen ermittelt.ErgebnisseDer Vergleich der Farbräume ergab keine relevant unterschiedlichen Korrektklassifikationsraten im Bereich von 69 % bis 72 %. Bei Zuschneidung auf die Lippen wurde mit 70 % eine deutlich höhere Korrektklassifikationsrate als bei Zuschnitt auf das gesamte Sprechergesicht (34 %) erreicht. Mit dem GRUConv-Modell betrugen die maximalen Korrektklassifikationsraten 87 % bei bekannten Sprechenden und 63 % in der Validierung mit unbekannten Sprechenden.SchlussfolgerungDas erstmals für die deutsche Sprache entwickelte neuronale Netzwerk zum Lippenlesen zeigt eine sehr große, mit englischsprachigen Algorithmen vergleichbare Genauigkeit. Es funktioniert auch mit unbekannten Sprechenden und kann mit mehr Wortklassen generalisiert werden.

Modelling Bathing Water Quality Using Official Monitoring Data

Predictive models of bathing water quality are a useful support to traditional monitoring and provide timely and adequate information for the protection of public health. When developing models, it is critical to select an appropriate model type and appropriate metrics to reduce errors so that the predicted outcome is reliable. It is usually necessary to conduct intensive sampling to collect a sufficient amount of data. This paper presents the process of developing a predictive model in Kaštela Bay (Adriatic Sea) using only data from regular (official) bathing water quality monitoring collected during five bathing seasons. The predictive modelling process, which included data preprocessing, model training, and model tuning, showed no silver bullet model and selected two model types that met the specified requirements: a neural network (ANN) for Escherichia coli and a random forest (RF) for intestinal enterococci. The different model types are probably the result of the different persistence of two indicator bacteria to the effects of marine environmental factors and consequently the different die-off rates. By combining these two models, the bathing water samples were classified with acceptable performances, an informedness of 71.7%, an F-score of 47.1%, and an overall accuracy of 80.6%.

Impact of climate change on hydro-climatological parameters in North Cyprus: application of artificial intelligence-based statistical downscaling models

Abstract There are many environmental challenges in water-limited places in the 21st century, particularly in dry and semi-arid regions, due to the threat of climate change caused by the greenhouse effect. This study intends to explore and assess the influence of climate change on hydro-climatological parameters using statistical downscaling and future forecasts of mean monthly precipitation and temperature throughout Famagusta (Mağusa), Nicosia (Lefkoşa), and Kyrenia (Girne) stations, North Cyprus. To achieve the study's goal, 13 predictors of BNU-ESM GCMs from CMIP5 were used at a grid point in the Karfas region. To find the primary predictors, GCM data were screened using mutual information (MI) and correlation coefficient (CC) feature extraction methods prior to downscaling modeling. A neural network (ANN), an adaptive neuro fuzzy inference system (ANFIS), and multiple linear regression (MLR) models were employed as the downscaling models. We used the best downscaling model as a benchmark for future precipitation and temperature estimates for the period 2018–2040 under the RCP4.5 scenario. In the future, Famagusta and Nicosia would have up to 22% less rain, and Famagusta and Kyrenia will have 2.9% greater heat. The findings of this research could be useful in decision-making, as well as water resource management and climate change.

Model Selection For Forecasting Rainfall Dataset

The objective of this research is to obtain the best method for forecasting rainfall in the Wonorejo reservoir in Surabaya. Time series and causal approaches using statistical methods and machine learning will be compared to forecast rainfall. Time series regression (TSR), autoregressive integrated moving average (ARIMA), linear regression (LR), and transfer function (TF) are used as a statistical method. Feedforward neural network (FFNN) and deep feed-forward neural network (DFFNN) is used as a machine learning method. Statistical methods are used to capture linear patterns, whereas the machine learning method is used to capture nonlinear patterns. Data about hourly rainfall in the Wonorejo reservoir is used as a case study. The data has a seasonal pattern, i.e. monthly seasonality. Based on the cross-validation and information criteria, the results showed that DFFNN using the time series approach has a more accurate forecast than other methods. In general, machine learning methods have better accuracy than statistical methods. Furthermore, additional information is obtained, through this research the parameter that best to make a neural network model is known. Moreover, these results are also not in line with the results of M3 and M4 competition, i.e. more complex methods do not necessarily produce better forecasts than simpler methods.

Prediction of Methanol Production in a Carbon Dioxide Hydrogenation Plant Using Neural Networks

The objective of this research was to design a neural network (ANN) to predict the methanol flux at the outlet of a carbon dioxide dehydrogenation plant. For the development of the ANN, a database was generated, in the open-source simulation software “DWSIM”, from the validation of a process described in the literature. The sample consists of 133 data pairs with four inputs: reactor pressure and temperature, mass flow of carbon dioxide and hydrogen, and one output: flow of methanol. The ANN was designed using 12 neurons in the hidden layer and it was trained with the Levenberg–Marquardt algorithm. In the training, validation and testing phase, a global mean square (RMSE) value of 0.0085 and a global regression coefficient R of 0.9442 were obtained. The network was validated through an analysis of variance (ANOVA), where the p-value for all cases was greater than 0.05, which indicates that there are no significant differences between the observations and those predicted by the ANN. Therefore, the designed ANN can be used to predict the methanol flow at the exit of a dehydrogenation plant and later for the optimization of the system.

Analysis of the Nosema Cells Identification for Microscopic Images.

The use of image processing tools, machine learning, and deep learning approaches has become very useful and robust in recent years. This paper introduces the detection of the Nosema disease, which is considered to be one of the most economically significant diseases today. This work shows a solution for recognizing and identifying Nosema cells between the other existing objects in the microscopic image. Two main strategies are examined. The first strategy uses image processing tools to extract the most valuable information and features from the dataset of microscopic images. Then, machine learning methods are applied, such as a neural network (ANN) and support vector machine (SVM) for detecting and classifying the Nosema disease cells. The second strategy explores deep learning and transfers learning. Several approaches were examined, including a convolutional neural network (CNN) classifier and several methods of transfer learning (AlexNet, VGG-16 and VGG-19), which were fine-tuned and applied to the object sub-images in order to identify the Nosema images from the other object images. The best accuracy was reached by the VGG-16 pre-trained neural network with 96.25%.

Probing Druggability of the Zika Virus Envelope Proteins using a Novel Computational Approach

A computational protocol is developed in order to search for potential drug binding sites in a protein, taking advantage of molecular dynamics (MD) simulation trajectories that account for thermal fluctuations of the system. The protocol combines several complementary methodologies, namely survey of cavities formed inside the protein of interest during an MD simulation trajectory as potential binding pockets, probing binding of diverse drug fragments to the protein, standard docking protocols, and scoring protein amino acids based on their conservation, intrinsic accessibility and druggability scores. We apply this protocol to investigate potential binding sites in a heterotetrameric subunit (E2M2) from the envelope of the Zika virus (ZIKV) as a target example. Notably, unlike dengue virus (DENV), potential drug binding sites are not well known for ZIKV. An equilibrated trajectory of E2M2 in solution is used as an input for four different independent potential drug binding site identification approaches. Specifically, Fpocket, FTMap, AutodockVina, and epitope scoring are used for narrowing drug binding site candidate search. Prediction of an effective drug takes advantage of flavivirus ChEMBL database (i.e., ZIKV and DENV) drugs for training a neural network (ANN) classification-model. Each drug in the training data set is preprocessed to (1) extract atom-bond node-edge information and molecular properties; (2) perform sequence-padding for a consistent dimension (3000 features); and (3) apply dimension reduction to reduce the high dimensional features to 2, without much loss of information. The preprocessed drug data input is then fed to ANN and the class and probability of ZIKV druggability are predicted. The predicted ZIKV drug is docked onto the identified binding site, and the complex is subjected to MD simulations to study the stability of the binding and associated energetics.

Sequence-based Identification of Allergen Proteins Developed by Integration of PseAAC and Statistical Moments via 5-Step Rule

Background: Allergens are antigens that can stimulate an atopic type I human hypersensitivity reaction by an immunoglobulin E (IgE) reaction. Some proteins are naturally allergenic than others. The challenge for toxicologists is to identify properties that allow proteins to cause allergic sensitization and allergic diseases. The identification of allergen proteins is a very critical and pivotal task. The experimental identification of protein functions is a hectic, laborious and costly task; therefore, computer scientists have proposed various methods in the field of computational biology and bioinformatics using various data science approaches. Objectives: Herein, we report a novel predictor for the identification of allergen proteins. Methods: For feature extraction, statistical moments and various position-based features have been incorporated into Chou’s pseudo amino acid composition (PseAAC), and are used for training of a neural network. Results: The predictor is validated through 10-fold cross-validation and Jackknife testing, which gave 99.43% and 99.87% accurate results. Conclusions: Thus, the proposed predictor can help in predicting the Allergen proteins in an efficient and accurate way and can provide baseline data for the discovery of new drugs and biomarkers.