Standard Backpropagation Algorithm Research Articles

Review of neuromorphic computing based on NAND flash memory.

The proliferation of data has facilitated global accessibility, which demands escalating amounts of power for data storage and processing purposes. In recent years, there has been a rise in research in the field of neuromorphic electronics, which draws inspiration from biological neurons and synapses. These electronics possess the ability to perform in-memory computing, which helps alleviate the limitations imposed by the 'von Neumann bottleneck' that exists between the memory and processor in the traditional von Neumann architecture. By leveraging their multi-bit non-volatility, characteristics that mimic biology, and Kirchhoff's law, neuromorphic electronics offer a promising solution to reduce the power consumption in processing vector-matrix multiplication tasks. Among all the existing nonvolatile memory technologies, NAND flash memory is one of the most competitive integrated solutions for the storage of large volumes of data. This work provides a comprehensive overview of the recent developments in neuromorphic computing based on NAND flash memory. Neuromorphic architectures using NAND flash memory for off-chip learning are presented with various quantization levels of input and weight. Next, neuromorphic architectures for on-chip learning are presented using standard backpropagation and feedback alignment algorithms. The array architecture, operation scheme, and electrical characteristics of NAND flash memory are discussed with a focus on the use of NAND flash memory in various neural network structures. Furthermore, the discrepancy of array architecture between on-chip learning and off-chip learning is addressed. This review article provides a foundation for understanding the neuromorphic computing based on the NAND flash memory and methods to utilize it based on application requirements.

Artificial Neural Network-Based Modeling of Performance Spark Ignition Engine Fuelled with Bioethanol and Gasoline

Machine learning technology can distinguish the relationship between engine characteristics and performances. Therefore, the goal of the present work is to predict the performance parameters of a single-cylinder 4-stroke gasoline engine at different ignition timings using a blended mixture of gasoline and bioethanol by an artificial neural network (ANN). Experimental data for training and testing in the proposed ANN was obtained at a dynamic speed and full load condition. An ANN model was developed based on standard Back-Propagation algorithm for the spark ignition engine. Multi-layer perception network (MLP) was used for non-linear mapping between the input and output parameters. An optimizer in the family of quasi-Newton methods (lbfgs) and the rectified linear unit function were used to assess the percentage error between the desired and the predicted values. The network input parameters are engine speed, fuel, and ignition timing. Furthermore, torque, power, specific fuel consumption (SFC), thermal efficiency (ηth), and energy consumption (EC) are taken as output parameters. The results show that ANN is the proper method for predicting SIE performance because it has accurate prediction results that are very similar to experimental results. Moreover, from the observation results, the ANN model can predict the engine performance quite well with correlation coefficient (R)=0.962139 and MSE=0.003967 for data testing.

Optimization of diesel engine performance and emission using waste plastic pyrolytic oil by ANN and its thermo-economic assessment.

The current study focuses on the engine performance and emission analysis of a 4-stroke compression ignition engine powered by waste plastic oil (WPO) obtained by the catalytic pyrolysis of medical plastic wastes. This is followed by their optimization study and economic analysis. This study demonstrates the use of artificial neural networks (ANN) to forecast a multi-component fuel mixture, which is novel and reduces the amount of experimental effort required to determine the engine output characteristics. The engine tests were conducted using WPO blended diesel at various proportions (10%, 20%, 30% by volume) to acquire the required data for training the ANN model, which enables better prediction for the engine performance by making use of the standard back-propagation algorithm. Considering supervised data obtained from repeated engine tests, an artificial intelligence-based model of ANN was designed to select different parameters of performance and emission as output layers; at the same time, engine loading and different blending ratios of the test fuels were taken as the input layers. The ANN model was built up making use of 80% of testing outcomes for training. The ANN model forecasted engine performance and exhaust emission with regression coefficients (R) at 0.989-0.998 intervals and a mean relative error from 0.002 to 0.348%. Such results illustrated the effectiveness of the ANN model for estimating emissions and the performance of diesel engines. Moreover, the economic viability of the use of 20WPO as an alternative to diesel was justified by thermo-economic analysis.

A shallow hybrid classical–quantum spiking feedforward neural network for noise-robust image classification

Deep Convolutional Neural Network (CNN)-based image classification systems are often susceptible to noise interruption, i.e., minor image noise may significantly impact the outcome. On the contrary, classical Spiking Neural Network (SNN) is known for handling noisy data due to the stochastic and temporal behaviour of the spiking neuron signals. However, it is not always feasible to train the weights of the classical SNN due to stochastic and non-differentiable spiking signals. Recent applications of quantum machine learning to stochastic modelling have predicted that quantum computing would prove to be a significant advantage, hence catapulting the study of quantum and neuromorphic computing to a new level of development. Motivated by these observations, this paper introduces a shallow hybrid classical–quantum spiking feedforward neural network referred to as Spiking Quantum Neural Network (SQNN) for dealing with a robust image classification task in the presence of noise and adversarial attacks. The proposed SQNN offers the inherent capabilities of processing unforeseen noisy test images due to its spatial and temporal information. An efficient variational quantum circuit training with the help of a standard back-propagation algorithm obviates the classical Spiking Time-Dependent Plasticity (STDP) and Spike-Prop algorithms, which are often found inefficient in training the feedforward SNN. The proposed SQNN is extensively tested and benchmarked on the PennyLane Quantum Simulator. Experimental results show that the proposed SQNN model supersedes the feedforward SNN (SFNN), Random Quantum Neural Networks (RQNN), AlexNet and ResNet-18 on unseen test images with added noise from the FashionMNIST, MNIST, KMNIST, CIFAR10, and ImageNet datasets.

A Tandem Learning Rule for Effective Training and Rapid Inference of Deep Spiking Neural Networks.

Spiking neural networks (SNNs) represent the most prominent biologically inspired computing model for neuromorphic computing (NC) architectures. However, due to the nondifferentiable nature of spiking neuronal functions, the standard error backpropagation algorithm is not directly applicable to SNNs. In this work, we propose a tandem learning framework that consists of an SNN and an artificial neural network (ANN) coupled through weight sharing. The ANN is an auxiliary structure that facilitates the error backpropagation for the training of the SNN at the spike-train level. To this end, we consider the spike count as the discrete neural representation in the SNN and design an ANN neuronal activation function that can effectively approximate the spike count of the coupled SNN. The proposed tandem learning rule demonstrates competitive pattern recognition and regression capabilities on both the conventional frame- and event-based vision datasets, with at least an order of magnitude reduced inference time and total synaptic operations over other state-of-the-art SNN implementations. Therefore, the proposed tandem learning rule offers a novel solution to training efficient, low latency, and high-accuracy deep SNNs with low computing resources.

A novel fault diagnosis scheme for a ballast system

The Back Propagation (BP) neural network model is one of the neural network models suitable for model recognition and classification. However, a single BP network model cannot fully tap the feature information hiding in the data and is vulnerable to the uncertainty associated with system parameter changes and also suffers from the effects of long diagnosis times, inaccurate and undesired diagnosis results in applying a large amount of high-dimensional feature data. In addition, the standard BP algorithm has limitations in pattern recognition and classification. For example, it adopts the quadratic cost function to train the neural network, which can lead to low convergence accuracy, low convergence rate, and so on. Compared with the quadratic cost function, the cross-entropy cost function can effectively improve the convergence accuracy and the convergence rate of the neural network. The Dempster–Shafer (D–S) evidence theory is a powerful tool for handling incomplete, imprecise, and uncertain information and merging it when it derives from multiple BP networks. In order to overcome the drawbacks of a single BP network, it is necessary to divide the high-dimensional feature data into several different categories of low-dimensional feature data and to input the same dimensional feature data into the same BP network model, as well as to design the information fusion algorithms based on D–S evidence theory and multiple improved BP networks with the cross-entropy cost function. For this reason, this paper proposes a novel information fusion diagnosis approach based on D–S evidence theory and multiple improved BP networks for fault diagnosis and classification of multiple devices in a ballast system subjected to extreme environmental conditions. The proposed diagnosis approach utilizes multiple neural networks to carry out local fault diagnosis, adopts the D–S evidence theory to fuse the results of local diagnosis and derive the decision results, and introduces an improved BP algorithm based on the cross-entropy cost function into the individual neural networks to enhance the convergence rate and diagnosis accuracy of the individual neural networks. This method can overcome the deficiencies of individual BP networks in multiple BP networks in their poor diagnostic ability to handle incomplete, imprecise, and uncertain sample information. According to the design idea of the proposed diagnosis approach, a new fusion modeling for fault diagnosis of multiple devices is constructed for a ballast system. The validity of the proposed diagnosis method and model is illustrated through both simulation studies and experimental tests.

Backpropagation Artificial Neural Network Enhancement using Beale-Powell Approach Technique

Machine learning algorithms can study existing data to perform specific tasks. One of the well-known machine learning algorithms is the backpropagation algorithm, but this algorithm often provides poor convergence speed in the training process and a long training time. The purpose of this study is to optimize the standard backpropagation algorithm using the Beale-Powell conjugate gradient algorithm so that the training time needed to achieve convergence is not too long, which later can be used as a reference and information for solving predictive problems. The Beale-Powell conjugate gradient algorithm can solve unlimited optimization problems and is much more efficient than gradient descent-based algorithms such as standard backpropagation. The research data used for the analysis were formal education participation data in Indonesia. To be trained and tested using the 7-10-1 architecture. The results showed that the Beale-Powell Conjugate Gradient algorithm could more quickly perform the training and convergence process. However, the MSE value of testing and performance is still superior to the backpropagation algorithm. So it can be concluded that for the prediction case of Formal Education Participation in Indonesia, the Conjugate Gradient Beale-Powell algorithm is good enough to optimize the performance of backpropagation standards seen from the convergence speed and training performance.

Deep neural networks for quick and precise geometry optimization of segmented thermoelectric generators

To solve the problems of the current optimization methods for solar segmented thermoelectric generator performance based on numerical methods, this paper applied deep neural networks to optimize the device geometry for improved thermo-mechanical performance. The motivation for using the deep neural network is to overcome the lengthy computational time and very high computational energy required by the traditional numerical method in optimizing the segmented thermoelectric generator performance. The numerical model is built using ANSYS software and the effects of temperature dependency in the 4 thermoelectric materials are considered to ensure result accuracy. Furthermore, 16 possible geometry parameters which were previously not considered, encompassing the individual and combined segment's heights and cross-sectional areas are optimized to find which set of parameters are the best in maximizing the device performance. The deep neural network is a regressive multilayer perceptron with network hyperparameters comprising 2 hidden layers with 5 neurons per layer. The training process is governed by the Levenberg-Marquardt standard backpropagation algorithm to minimize the mean squared error and maximize the regression correlation between the neural network forecasted outputs and the numerical-generated dataset. The most significant contribution of the proposed deep neural network is that it was able to quickly and accurately forecast the device performance in just 10 s, which was 2880 times faster than the conventional numerical-based optimization approach. Additionally, the optimized device had a maximum efficiency of 18%, which was 78% higher than that of the unoptimized device. Also, the thermal stress of the optimized device was 73% less than that of the unoptimized device design, indicating an extension in the device mechanical reliability and service lifetime. The results reported in this paper will accelerate the ease at which efficient, long-lasting segmented thermoelectric generators are manufactured by harnessing the power of artificial intelligence.

Spatiality Sensitive Learning for Cancer Metastasis Detection in Whole-Slide Images

Metastasis detection in lymph nodes via microscopic examination of histopathological images is one of the most crucial diagnostic procedures for breast cancer staging. The manual analysis is extremely labor-intensive and time-consuming because of complexities and diversities of histopathology images. Deep learning has been utilized in automatic cancer metastasis detection in recent years. Due to the huge size of whole-slide images, most existing approaches split each image into smaller patches and simply treat these patches independently, which ignores the spatial correlations among them. To solve this problem, this paper proposes an effective spatially sensitive learning framework for cancer metastasis detection in whole-slide images. Moreover, a novel spatial loss function is designed to ensure the consistency of prediction over neighboring patches. Specifically, through incorporating long short-term memory and spatial loss constraint on top of a convolutional neural network feature extractor, the proposed method can effectively learn both the appearance of each patch and spatial relationships between adjacent image patches. With the standard back-propagation algorithm, the whole framework can be trained in an end-to-end way. Finally, the regions with high tumor probability in the resulting probability map are the metastasis locations. Extensive experiments on the benchmark Camelyon 2016 Grand Challenge dataset show the effectiveness of the proposed approach with respect to state-of-the-art competitors. The obtained precision, recall, and balanced accuracy are 0.9565, 0.9167, and 0.9458, respectively. It is also demonstrated that the proposed approach can provide more accurate detection results and is helpful for early diagnosis of cancer metastasis.

Successfully and efficiently training deep multi-layer perceptrons with logistic activation function simply requires initializing the weights with an appropriate negative mean

The vanishing gradient problem (i.e., gradients prematurely becoming extremely small during training, thereby effectively preventing a network from learning) is a long-standing obstacle to the training of deep neural networks using sigmoid activation functions when using the standard back-propagation algorithm. In this paper, we found that an important contributor to the problem is weight initialization.We started by developing a simple theoretical model showing how the expected value of gradients is affected by the mean of the initial weights. We then developed a second theoretical model that allowed us to identify a sufficient condition for the vanishing gradient problem to occur. Using these theories we found that initial back-propagation gradients do not vanish if the mean of the initial weights is negative and inversely proportional to the number of neurons in a layer.Numerous experiments with networks with 10 and 15 hidden layers corroborated the theoretical predictions: If we initialized weights as indicated by the theory, the standard back-propagation algorithm was both highly successful and efficient at training deep neural networks using sigmoid activation functions.

Backpropagation with BFGS Optimizer for Covid-19 Prediction Cases in Surabaya

Covid-19 is a new type of corona virus called SARS-CoV-2. One of the cities that has contributed the most to infected Covid-19 cases in Indonesia is Surabaya, East Java. Predicting the Covid-19 is the important thing to do. One of the prediction methods is Artificial Neural Network (ANN). The backpropagation algorithm is one of the ANN methods that has been successfully used in various fields. However, the performance of backpropagation is depended on the architecture and optimization method. The standard backpropagation algorithm is optimized by gradient descent method. The Broyden - Fletcher - Goldfarb - Shanno (BFGS) algorithm works faster then gradient descent. This paper was predicting the Covid-19 cases in Surabaya using backpropagation with BFGS. Several scenarios of backpropagation parameters were also tested to produce optimal performance. The proposed method gives better results with a faster convergence then the standard backpropagation algorithm for predicting the Covid-19 cases in Surabaya.

Artificial neural networks for cosmic gamma-ray propagation in the universe

We explore the potential of an artificial neural network (ANN) based method to probe the propagation of cosmic γ-ray photons in the extragalactic Universe. The journey of γ-rays emitted from a distant source like blazar to the observer at the Earth is impeded by the absorption through the interaction with the extragalactic background light (EBL), leading to an electron-positron pair production. This process dominates for gamma ray photons with energy above 10 GeV propagating over the cosmological distances. The effect of γ-ray attenuation is characterized by a physical quantity called optical depth, which strongly depends on the γ-ray photon energy, redshift of the source, and density of the EBL photons. We estimate the optical depth values for γ-ray energies above 10 GeV emitted from the sources at redshifts in the range 0.01 to 1 using three different and most promising EBL models. These optical depth estimates are randomly divided into two data sets for training and testing of the ANN using energy, redshift as inputs and optical depth as output. The optimization of ANN-performance for each EBL model employs standard back-propagation (BP) and radial-basis function (RBF) algorithms. The performance of the ANN model using the RBF is found to be superior to the BP method. In particular, the RBF-ANN with 40 neurons in the hidden layer corresponding to the EBL model proposed by Finke et al. (2010) shows the best performance for the propagation of γ-rays in the Universe.

Attention-Based CNN-RNN Arabic Text Recognition from Natural Scene Images

According to statistics, there are 422 million speakers of the Arabic language. Islam is the second-largest religion in the world, and its followers constitute approximately 25% of the world’s population. Since the Holy Quran is in Arabic, nearly all Muslims understand the Arabic language per some analytical information. Many countries have Arabic as their native and official language as well. In recent years, the number of internet users speaking the Arabic language has been increased, but there is very little work on it due to some complications. It is challenging to build a robust recognition system (RS) for cursive nature languages such as Arabic. These challenges become more complex if there are variations in text size, fonts, colors, orientation, lighting conditions, noise within a dataset, etc. To deal with them, deep learning models show noticeable results on data modeling and can handle large datasets. Convolutional neural networks (CNNs) and recurrent neural networks (RNNs) can select good features and follow the sequential data learning technique. These two neural networks offer impressive results in many research areas such as text recognition, voice recognition, several tasks of Natural Language Processing (NLP), and others. This paper presents a CNN-RNN model with an attention mechanism for Arabic image text recognition. The model takes an input image and generates feature sequences through a CNN. These sequences are transferred to a bidirectional RNN to obtain feature sequences in order. The bidirectional RNN can miss some preprocessing of text segmentation. Therefore, a bidirectional RNN with an attention mechanism is used to generate output, enabling the model to select relevant information from the feature sequences. An attention mechanism implements end-to-end training through a standard backpropagation algorithm.

Service quality evaluation of bus lines based on improved momentum back‐propagation neural network model: A study of Hangzhou in China

This study was focused on Hangzhou in China that are undergoing large-scale subway construction, and an improved momentum back-propagation (BP) neural network model was trained. The model can analyze the complex traffic data, evaluate the service quality of bus line, and improve the estimation accuracy and convergence speed. For the same training data set, the convergence time of the BP algorithm with momentum term is reduced by 0.043 secs, the iterative convergence speed is improved by 0.66%, and the estimation accuracy is improved by 26.7% compared with the standard BP algorithm. Under similar conditions, the convergence time is 1.562 secs less than that of the standard BP algorithm, and the convergence speed was 24.1% higher than that of the standard BP algorithm, and the absolute value of the estimated error was less than 1%. Finally, a representative bus line in Hangzhou was used as an example to evaluate the model. The results showed that the improved momentum BP neural network model had a faster convergence speed and higher prediction accuracy of the comprehensive weight of bus line service quality. The prediction results of the model are consistent with the actual survey results, which indicates that the model constructed is reasonable.

Proposed Back Propagation Deep Neural Network for Intrusion Detection in Internet of Things Fog Computing

Internet of things (IoT) is an emerging concept which aims to connect billions of devices with each other anytime regardless of their location. Sadly, these IoT devices do not have enough computing resources to process huge amount of data. Therefore, Cloud computing is relied on to provide these resources. However, cloud computing based architecture fails in applications that demand very low and predictable latency, therefore the need for fog computing which is a new paradigm that is regarded as an extension of cloud computing to provide services between end users and the cloud user. Unfortunately, Fog-IoT is confronted with various security and privacy risks and prone to several cyberattacks which is a serious challenge. The purpose of this work is to present security and privacy threats towards Fog-IoT platform and discuss the security and privacy requirements in fog computing. We then proceed to propose an Intrusion Detection System (IDS) model using Standard Deep Neural Network's Back Propagation algorithm (BPDNN) to mitigate intrusions that attack Fog-IoT platform. The experimental Dataset for the proposed model is obtained from the Canadian Institute for Cybersecurity 2017 Dataset. Each instance of the attack in the dataset is separated into separate files, which are DoS (Denial of Service), DDoS (Distributed Denial of Service), Web Attack, Brute Force FTP, Brute Force SSH, Heartbleed, Infiltration and Botnet (Bot Network) Attack. The proposed model is trained using a 3-layer BP-DNN

Prediction of Chemical Oxygen Demand from The Chemical Composition of Wastewater by Artificial Neural Networks

In our era, many technical applications are being used. Artificial Neural Networks (ANNs) as one of the artificial intelligence tools have emerged to learn and discover a model of dynamic nonlinear. In this study, six input parameters were taken to predict the value of the Chemical Oxygen Demand (COD) in the wastewater before and after the treatment at the North Gas Company/Kirkuk, by using the standard back propagation algorithm. The network was trained with the 150 data collected from the quality indices of the untreated and treated wastewater, such as total chloride ions Cl-, nitrate ions NO3 −, phosphate ions PO4 −3, sulfate ions SO4 −2, ammonia NH3, Biochemical Oxygen Demand (BOD5) to predict one element, that is the COD. After properly training of the neural network, it was tested by using the test data, and the best results were selected by the consideration of the mean square error and the regression coefficient, where the best result appeared before wastewater treatment is 0.98235 and the best result after wastewater treatment is 0.99999. The findings of this study suggest that artificial neural networks are accurate and effective tools for predicting the COD values of treated wastewater.

Artificial Intelligence‐Based Surface Roughness Estimation Modelling for Milling of AA6061 Alloy

This study introduces the improvement of mathematical and predictive models of surface roughness parameter (Ra) in milling AA6061 alloy using carbide cutting tools coated with CVD‐TiCN in dry condition. An experimental model has been improved for estimating the surface roughness using artificial neural networks (ANN) and response surface methodology (RSM). For these models, cutting speed, depth of cut, and feed rate were evaluated as input parameters for experimental design. For the ANN modelling, the standard backpropagation algorithm was established to be the optimum selection for training the model. In the forming of the network construction, five different learning algorithms were used: the conjugate gradient backpropagation, Levenberg–Marquardt, scaled conjugate gradient, quasi‐Newton backpropagation, and resilient backpropagation. The best consequent with single hidden layers for the surface roughness was obtained by 3‐8‐1 network structures. The statistical analysis was performed with RSM‐based second‐order mathematics model. The influences of the cutting parameters on surface roughness were defined by using analysis of variance (ANOVA). The ANOVA results show that the depth of cut is the most effective parameter on surface roughness. Prediction models developed using ANN and RSM were compared in terms of prediction accuracy R2, MEP, and RMSE. The data estimated from ANN and RSM were realized to be very close to the data acquired from experimental studies. The value R2 of RSM model was higher than the values of the ANN model which demonstrated the stability and sturdiness of the RSM method.

Solving Riccati-Type Nonlinear Differential Equations with Novel Artificial Neural Networks

In this study we investigate neural network solutions to nonlinear differential equations of Ricatti-type. We employ a feed-forward Multilayer Perceptron Neural Network (MLPNN), but avoid the standard back-propagation algorithm for updating the intrinsic weights. Our objective is to minimize an error, which is a function of the network parameters i.e., the weights and biases. Once the weights of the neural network are obtained by our systematic procedure, we need not adjust all the parameters in the network, as postulated by many researchers before us, in order to achieve convergence. We only need to fine-tune our biases which are fixed to lie in a certain given range, and convergence to a solution with an acceptable minimum error is achieved. This greatly reduces the computational complexity of the given problem. We provide two important ODE examples, the first is a Ricatti type differential equation to which the procedure is applied, and this gave us perfect agreement with the exact solution. The second example however provided us with only an acceptable approximation to the exact solution. Our novel artificial neural networks procedure has demonstrated quite clearly the function approximation capabilities of ANN in the solution of nonlinear differential equations of Ricatti type.

A Novel Method for Solving Ordinary Differential Equations with Artificial Neural Networks

This research work investigates the use of Artificial Neural Network (ANN) based on models for solving first and second order linear constant coefficient ordinary differential equations with initial conditions. In particular, we employ a feed-forward Multilayer Perceptron Neural Network (MLPNN), but bypass the standard back-propagation algorithm for updating the intrinsic weights. A trial solution of the differential equation is written as a sum of two parts. The first part satisfies the initial or boundary conditions and contains no adjustable parameters. The second part involves a feed-forward neural network to be trained to satisfy the differential equation. Numerous works have appeared in recent times regarding the solution of differential equations using ANN, however majority of these employed a single hidden layer perceptron model, incorporating a back-propagation algorithm for weight updation. For the homogeneous case, we assume a solution in exponential form and compute a polynomial approximation using statistical regression. From here we pick the unknown coefficients as the weights from input layer to hidden layer of the associated neural network trial solution. To get the weights from hidden layer to the output layer, we form algebraic equations incorporating the default sign of the differential equations. We then apply the Gaussian Radial Basis function (GRBF) approximation model to achieve our objective. The weights obtained in this manner need not be adjusted. We proceed to develop a Neural Network algorithm using MathCAD software, which enables us to slightly adjust the intrinsic biases. We compare the convergence and the accuracy of our results with analytic solutions, as well as well-known numerical methods and obtain satisfactory results for our example ODE problems.

A New Initialization Approach in Particle Swarm Optimization for Global Optimization Problems.

Particle swarm optimization (PSO) algorithm is a population-based intelligent stochastic search technique used to search for food with the intrinsic manner of bee swarming. PSO is widely used to solve the diverse problems of optimization. Initialization of population is a critical factor in the PSO algorithm, which considerably influences the diversity and convergence during the process of PSO. Quasirandom sequences are useful for initializing the population to improve the diversity and convergence, rather than applying the random distribution for initialization. The performance of PSO is expanded in this paper to make it appropriate for the optimization problem by introducing a new initialization technique named WELL with the help of low-discrepancy sequence. To solve the optimization problems in large-dimensional search spaces, the proposed solution is termed as WE-PSO. The suggested solution has been verified on fifteen well-known unimodal and multimodal benchmark test problems extensively used in the literature, Moreover, the performance of WE-PSO is compared with the standard PSO and two other initialization approaches Sobol-based PSO (SO-PSO) and Halton-based PSO (H-PSO). The findings indicate that WE-PSO is better than the standard multimodal problem-solving techniques. The results validate the efficacy and effectiveness of our approach. In comparison, the proposed approach is used for artificial neural network (ANN) learning and contrasted to the standard backpropagation algorithm, standard PSO, H-PSO, and SO-PSO, respectively. The results of our technique has a higher accuracy score and outperforms traditional methods. Also, the outcome of our work presents an insight on how the proposed initialization technique has a high effect on the quality of cost function, integration, and diversity aspects.