Search results

New approaches for non‐classical neural‐based computing are introduced. The developed approaches utilize new concepts in three‐dimensionality, invertibility and reversibility to perform the required neural computing. The various implementations of the new neural circuits using the introduced paradigms and architectures are presented, several applications are shown, and the extension for the utilization in neural‐systolic computing is also introduced.

The new neural paradigms utilize new findings in computational intelligence and advanced logic synthesis to perform the functionality of the basic neural network (NN). This includes the techniques of three‐dimensionality, invertibility and reversibility. The extension of implementation to neural‐systolic computing using the introduced reversible neural‐systolic architecture is also presented.

Novel NN paradigms are introduced in this paper. New 3D paradigm of NL circuits called three‐dimensional inverted neural logic (3DINL) circuits is introduced. The new 3D architecture inverts the inputs and weights in the standard neural architecture: inputs become bases on internal interconnects, and weights become leaves of the network. New reversible neural network (RevNN) architecture is also introduced, and a RevNN paradigm using supervised learning is presented. The applications of RevNN to multiple‐output feedforward discrete plant control and to reversible neural‐systolic computing are also shown. Reversible neural paradigm that includes reversible neural architecture utilizing the extended mapping technique with an application to the reversible solution of the maze problem using the reversible counterpropagation NN is introduced, and new neural paradigm of reversibility in both architecture and training using reversibility in independent component analysis is also presented.

Since the new 3D NNs can be useful as a possible optimal design choice for compacting a learning (trainable) circuit in 3D space, and because reversibility is essential in the minimal‐power computing as the reduction of power consumption is a main requirement for the circuit synthesis of several emerging technologies, the introduced methods for non‐classical neural computation are new and interesting for the design of several future technologies that require optimal design specifications such as three‐dimensionality, regularity, super‐high speed, minimum power consumption and minimum size such as in low‐power control, adiabatic signal processing, quantum computing, and nanotechnology.

The purpose of this paper is to obtain an estimation of not measured mechanical state variables of the drive system with elastic coupling between the driven motor and a load machine, using neural networks (NN) of different type for the sensorless drive system.

The load‐side speed and the torsional torque are estimated using multi‐layer perceptron (MLP) and radial basis function (RBF) networks. The special forms of input vectors for neural state estimators were proposed and tested in open‐ and closed‐loop control structure. The estimation quality as well as sensitivity of neural estimators to the changes of the inertia moment of the load machine were evaluated and compared.

It is shown that an application of RBF‐based neural estimators can give better accuracy of the load speed and torsional torque estimation, especially for the proper choice of the input vector of NN, also in the case of a big change of the load machine time constant.

The investigation and comparison is based on simulation tests and looked mainly at the quality of state variable estimation while the realisation cost in parallel processing devices (FPGA) still need to be addressed.

The proposed neural state variable estimators of two‐mass system can be practically implemented in the control structure of two‐mass drive with additional feedbacks from load machine speed and torsional torque, which results in the successive vibration damping.

The application of RBF neural state estimators for two‐mass drive and their comparison with commonly used MLP‐based estimators, as well as testing of both type of NN in the closed‐loop control structure with additional feedbacks based on state variables estimated by neural estimators.

In this research, neural network (NN) and genetic algorithm (GA) are used together to design optimal NN structure. The proposed approach combines the characteristics of GA and NN to reduce the computational complexity of artificial intelligence applications in design and manufacturing. Genetic input selection approach is introduced to obtain optimal NN topology. Experimental results are given to evaluate the performance of the proposed system.

The purpose of this paper is to develop an efficient brain tumor detection model using the beneficial concept of hybrid classification using magnetic resonance imaging (MRI) images Brain tumors are the most familiar and destructive disease, resulting to a very short life expectancy in their highest grade. The knowledge and the sudden progression in the area of brain imaging technologies have perpetually ready for an essential role in evaluating and concentrating the novel perceptions of brain anatomy and operations. The system of image processing has prevalent usage in the part of medical science for enhancing the early diagnosis and treatment phases.

The proposed detection model involves five main phases, namely, image pre-processing, tumor segmentation, feature extraction, third-level discrete wavelet transform (DWT) extraction and detection. Initially, the input MRI image is subjected to pre-processing using different steps called image scaling, entropy-based trilateral filtering and skull stripping. Image scaling is used to resize the image, entropy-based trilateral filtering extends to eradicate the noise from the digital image. Moreover, skull stripping is done by Otsu thresholding. Next to the pre-processing, tumor segmentation is performed by the fuzzy centroid-based region growing algorithm. Once the tumor is segmented from the input MRI image, feature extraction is done, which focuses on the first-order and higher-order statistical measures. In the detection side, a hybrid classifier with the merging of neural network (NN) and convolutional neural network (CNN) is adopted. Here, NN takes the first-order and higher-order statistical measures as input, whereas CNN takes the third level DWT image as input. As an improvement, the number of hidden neurons of both NN and CNN is optimized by a novel meta-heuristic algorithm called Crossover Operated Rooster-based Chicken Swarm Optimization (COR-CSO). The AND operation of outcomes obtained from both optimized NN and CNN categorizes the input image into two classes such as normal and abnormal. Finally, a valuable performance evaluation will prove that the performance of the proposed model is quite good over the entire existing model.

From the experimental results, the accuracy of the suggested COR-CSO-NN + CNN was seemed to be 18% superior to support vector machine, 11.3% superior to NN, 22.9% superior to deep belief network, 15.6% superior to CNN and 13.4% superior to NN + CNN, 11.3% superior to particle swarm optimization-NN + CNN, 9.2% superior to grey wolf optimization-NN + CNN, 5.3% superior to whale optimization algorithm-NN + CNN and 3.5% superior to CSO-NN + CNN. Finally, it was concluded that the suggested model is superior in detecting brain tumors effectively using MRI images.

This paper adopts the latest optimization algorithm called COR-CSO to detect brain tumors using NN and CNN. This is the first study that uses COR-CSO-based optimization for accurate brain tumor detection.

Neural Network (NN) simulation models are being increasingly utilised in the business and management fields as forecasting, pattern recognition and classification tools. Their growing popularity appears to emanate from the ability of NNs to approximate complex non‐linear relationships, via their capacity to represent latent combinations of unobservable variables in hidden layers. Although there is a growing business literature on the ability of NNs to predict various corporate outcomes (e.g., corporate failure), and to forecast time series data (e.g., share prices), they have yet to be fully evaluated by business academics on cross‐sectional data. This paper provides an overview of the NN modelling approach and compares the performance of NNs, relative to conventional OLS regression analysis, in predicting the cross‐sectional variation in corporate audit fees. The empirical results suggest that the NN models exhibit superior forecasting accuracy to their OLS counterparts, but that this differential reduces when the models are tested out‐of‐sample.

In this paper we are presenting our research findings on how effective neural networks are at forecasting and estimating preliminary project costs. We have shown that neural networks completely outperform traditional techniques in such tasks. In exploring nonlinear techniques almost all of the current research involves neural network techniques, especially multilayer perceptron (MLP) models and other statistical techniques and few authors have considered radial basis function neural network (RBF NN) models in their research. For this purpose we have developed RBF NN models to represent nonlinear static and dynamic processes and compared their performance with traditional methods. The traditional methods applied in this paper are multiple linear regression (MLR) and autoregressive moving average models with eXogenous input (ARMAX). The performance of these and RBF neural network and traditional models is tested on common data sets and their results are presented.

The purpose of this paper is to develop e-health and patient monitoring systems remotely to overcome the difficulty of patients going to hospitals especially in times of epidemics such as virus disease (COVID-19). Artificial intelligence (AI) technology will be combined Internet of Things (IoT) in this research to overcome these challenges. The research aims to select the most appropriate, best-hidden layers numbers and the activation function types for the neural network (NN). Then, define the patient data sent through protocols of the IoT. NN checks the patient’s medical sensors data to make the appropriate decision. Then it sends this diagnosis to the doctor. Using the proposed solution, the patients can diagnose and expect the disease automatically and help physicians to discover and analyze the disease remotely without the need for patients to go to the hospital.

AI technology will be combined with the IoT in this research. The research aims to select the most appropriate’ best-hidden layers numbers’ and the activation function types for the NN.

Decision support health-care system based on IoT and deep learning techniques was proposed. The authors checked out the ability to integrate the deep learning technique in the automatic diagnosis and IoT abilities for speeding message communication over the internet has been investigated in the proposed system. The authors have chosen the appropriate structure of the NN (best-hidden layers numbers and the activation function types) to build the e-health system is performed in this work. Also, depended on the data from expert physicians to learn the NN in the e-health system. In the verification mode, the overall evaluation of the proposed diagnosis health-care system gives reliability under different patient’s conditions. From evaluation and simulation results, it is clear that the double hidden layer of feed-forward NN and its neurons contain Tanh function preferable than other NN.

AI technology will be combined IoT in this research to overcome challenges. The research aims to select the most appropriate, best-hidden layers numbers and the activation function types for the NN.

Diabetic retinopathy (DR) is a central root of blindness all over the world. Though DR is tough to diagnose in starting stages, and the detection procedure might be time-consuming even for qualified experts. Nowadays, intelligent disease detection techniques are extremely acceptable for progress analysis and recognition of various diseases. Therefore, a computer-aided diagnosis scheme based on intelligent learning approaches is intended to propose for diagnosing DR effectively using a benchmark dataset.

The proposed DR diagnostic procedure involves four main steps: (1) image pre-processing, (2) blood vessel segmentation, (3) feature extraction, and (4) classification. Initially, the retinal fundus image is taken for pre-processing with the help of Contrast Limited Adaptive Histogram Equalization (CLAHE) and average filter. In the next step, the blood vessel segmentation is carried out using a segmentation process with optimized gray-level thresholding. Once the blood vessels are extracted, feature extraction is done, using Local Binary Pattern (LBP), Texture Energy Measurement (TEM based on Laws of Texture Energy), and two entropy computations – Shanon's entropy, and Kapur's entropy. These collected features are subjected to a classifier called Neural Network (NN) with an optimized training algorithm. Both the gray-level thresholding and NN is enhanced by the Modified Levy Updated-Dragonfly Algorithm (MLU-DA), which operates to maximize the segmentation accuracy and to reduce the error difference between the predicted and actual outcomes of the NN. Finally, this classification error can correctly prove the efficiency of the proposed DR detection model.

The overall accuracy of the proposed MLU-DA was 16.6% superior to conventional classifiers, and the precision of the developed MLU-DA was 22% better than LM-NN, 16.6% better than PSO-NN, GWO-NN, and DA-NN. Finally, it is concluded that the implemented MLU-DA outperformed state-of-the-art algorithms in detecting DR.

This paper adopts the latest optimization algorithm called MLU-DA-Neural Network with optimal gray-level thresholding for detecting diabetic retinopathy disease. This is the first work utilizes MLU-DA-based Neural Network for computer-aided Diabetic Retinopathy diagnosis.

The purpose of this paper is to introduce a neural network‐based market agent, which develops optimal bidding strategies for a power generating company (Genco) in a day‐ahead electricity market.

The problem of finding optimal bidding strategy for a Genco is formulated as a two‐level optimization problem. At the top level, the Genco aims at maximizing its total daily profit, and at the bottom level, the independent system operator obtains the power dispatch quantity for each market participant with the objective of maximizing the social welfare. The neural network is trained using a particle swarm optimization (PSO) algorithm with the objective of maximizing daily profit for the Genco.

The effectiveness of the proposed approach is established through several case studies on the benchmark IEEE 30‐bus test system for the day‐ahead market, with an hourly clearing mechanism and dynamically changing demand profile. Both block bidding and linear supply function bidding are considered for the Gencos and the variation of optimal bidding strategy with the change in demand is investigated. The performance is also evaluated in the context of the Brazilian electricity market with real market data and compared with the other methods reported in the literature.

Strategic bidding is a peculiar phenomenon observed in an oligopolistic electricity market and has several implications on policy making and mechanism design. In this work, the transmission line constraints and demand side bidding are taken into account for a more realistic simulation.

To the best of the authors' knowledge, this paper has introduced, for the first time, a neural network‐based market agent to develop optimal bidding strategies of a Genco in an electricity market. Simulation results obtained from the IEEE 30‐bus test system and the Brazilian electricity market demonstrate the superiority of the proposed approach, as compared to the conventional PSO‐based method and the genetic fuzzy rule‐based system approach, respectively.