Weights Of Artificial Neural Network Research Articles

Untrained neural networks can demonstrate memorization-independent abstract reasoning

The nature of abstract reasoning is a matter of debate. Modern artificial neural network (ANN) models, like large language models, demonstrate impressive success when tested on abstract reasoning problems. However, it has been argued that their success reflects some form of memorization of similar problems (data contamination) rather than a general-purpose abstract reasoning capability. This concern is supported by evidence of brittleness, and the requirement of extensive training. In our study, we explored whether abstract reasoning can be achieved using the toolbox of ANNs, without prior training. Specifically, we studied an ANN model in which the weights of a naive network are optimized during the solution of the problem, using the problem data itself, rather than any prior knowledge. We tested this modeling approach on visual reasoning problems and found that it performs relatively well. Crucially, this success does not rely on memorization of similar problems. We further suggest an explanation of how it works. Finally, as problem solving is performed by changing the ANN weights, we explored the connection between problem solving and the accumulation of knowledge in the ANNs.

Intelligent frequency control of AC microgrids with communication delay: An online tuning method subject to stabilizing parameters

Smart control techniques have been implemented to address fluctuating power levels within isolated microgrids, mitigating the risk of unstable frequencies and the potential degradation of power supply quality. However, a challenge lies in the fact that employing these computationally complex methods without stability preservation might not suffice to handle the rapid changes of this highly dynamic environment in real-world scenarios over communication delays. This study introduces a flexible real-time approach for the frequency control problem using an artificial neural network (ANN) constrained to stabilized regions. Our solution integrates stabilizing PID controllers, computed through small-signal analysis and tuned via an automated search for optimal ANN weights and reinforcement learning (RL)-based selected constraints. First, we design stabilizing PID controllers by applying the stability boundary locus method and the Mikhailov criterion, specifically addressing communication delays. Next, we refine the controller parameters online through an automated process that identifies optimal coefficient combinations, leveraging a constrained ANN to manage frequency deviations within a restricted parameter range. Our approach is further enhanced by employing the RL technique, which trains the tuning system using an interpolated stability boundary curve to ensure both stability and performance. This one-of-a-kind combination of ANN, RL, and advanced PID tuning methods is a big step forward in how we handle frequency control problems in isolated AC microgrids. The experiments show that our solution outperforms traditional methods due to its reduced parameter search space. In particular, the proposed method reduces transient and steady-state frequency deviations more than semi- and unconstrained methods. The improved metrics and stability analysis show that the method improves system performance and stability under changing conditions.

Tunable magnetic synapse for reliable neuromorphic computing

Artificial neural networks (ANNs), inspired by the structure and function of the human brain, have achieved remarkable success in various fields. However, ANNs implemented using conventional complementary metal oxide semiconductor technology face significant limitations. This has prompted exploration of nonvolatile memory technologies as potential solutions to overcome these limitations by integrating storage and computation within a single device. These emerging technologies can retain resistance values without power, allowing them to serve as analog weights in ANNs, mimicking the behavior of biological synapses. While promising, these nonvolatile devices often exhibit inherent nonlinear relationships between resistance and applied voltage, complicating training processes and potentially impacting learning accuracy. This article proposes a magnetic synapse device based on the spin–orbit torque effect with geometrically controlled linear and nonlinear response characteristics. The device consists of a magnetic multilayer stack patterned into a designed shape, where the width variation along the current flow direction allows for controllable magnetic domain wall propagation. Through finite element method simulations and experimental studies, we demonstrate that by engineering the device geometry, a linear relationship between the applied current and the resulting Hall resistance can be achieved, mimicking the desired linear weight-input behavior in artificial neural networks. Additionally, this study explores the influence of current pulse width on the response curves, revealing a deviation from linearity at longer pulse durations. The geometric tunability of the magnetic synapse device offers a promising approach for realizing reliable and energy-efficient neuromorphic computing architectures.

Thermophysical dynamics of magnetohydrodynamic electroosmotic two phase mixed convective flow in vertical annulus: neuro computation

ABSTRACT The thermophysical dynamics of electro – magnetohydrodynamic mixed convective electrically conducting fluid flow in the vertical cylindrical annulus are explored with electric and magnetic fields. The flow and heat transfer properties are explicitly examined in viscous fluid and electroosmotic flow regions. The coupling between an external pressure gradient, electroosmosis, and an extended electromagnetic field can control fluid flow and heat transfer. The governing equations associated with the physical phenomenon are addressed with the Galerkin finite element method. Eloquently, the analysis revealed that, in the absence of a lateral electric field, the flow velocity decelerates as the Hartmann number increases, which in turn ultimately, causes the Nusselt number to rise. Since a lateral electric field is involved, the flow field, temperature field, and Nusselt number profiles are presented in two regions. The artificial neural network model discusses the thermal transport phenomenon in both regions. The artificial neural network models regression, performance plots, weights, and bias are presented. The research outcome can be utilized to design exquisite and efficacious electromagnetic devices, particularly within a particular range of thermal physical properties.

A memetic dynamic coral reef optimisation algorithm for simultaneous training, design, and optimisation of artificial neural networks

Artificial Neural Networks (ANNs) have been used in a multitude of real-world applications given their predictive capabilities, and algorithms based on gradient descent, such as Backpropagation (BP) and variants, are usually considered for their optimisation. However, these algorithms have been shown to get stuck at local optima, and they require a cautious design of the architecture of the model. This paper proposes a novel memetic training method for simultaneously learning the ANNs structure and weights based on the Coral Reef Optimisation algorithms (CROs), a global-search metaheuristic based on corals’ biology and coral reef formation. Three versions based on the original CRO combined with a Local Search procedure are developed: (1) the basic one, called Memetic CRO; (2) a statistically guided version called Memetic SCRO (M-SCRO) that adjusts the algorithm parameters based on the population fitness; (3) and, finally, an improved Dynamic Statistically-driven version called Memetic Dynamic SCRO (M-DSCRO). M-DSCRO is designed with the idea of improving the M-SCRO version in the evolutionary process, evaluating whether the fitness distribution of the population of ANNs is normal to automatically decide the statistic to be used for assigning the algorithm parameters. Furthermore, all algorithms are adapted to the design of ANNs by means of the most suitable operators. The performance of the different algorithms is evaluated with 40 classification datasets, showing that the proposed M-DSCRO algorithm outperforms the other two versions on most of the datasets. In the final analysis, M-DSCRO is compared against four state-of-the-art methods, demonstrating its superior efficacy in terms of overall accuracy and minority class performance.

2:1 Multiplexer, 1:2 De-multiplexer,2:4 Decoder and 4:2 Encoder Circuit Design with CMOS Technology Implementing with Artificial Neural Network with Verilog HDL Code for Output

The goal of this project is to create a De-multiplexer gate using a complementary metal oxide semiconductor (CMOS) and an artificial neural network.When designing any COMS circuit, we always keep in mind that the lowest possible cost should be the aim.In this work, the circuit was designed using a multilayer artificial neural network. We utilize weights to alter the value and treat negative values as inverters and neurons as transistors in our work.We are also developing Verilog-HdL code to easily apply the De-multiplexer for experimenting with an artificial neural network and set weights to acquire the desired results. The purpose of this project is to use an artificial neural network and a complementary metal oxide semiconductor (CMOS) to develop a decoder gate.We constantly keep in mind that the goal should be to construct a COMS circuit at the lowest feasible cost.An artificial neural network with multiple layers was used to create the circuit in this work. In our study, we treat negative values as inverters and neurons as transistors, and we modify the value using weights.Additionally, we are creating Verilog-HdL code that will make it simple to use the Decoder to experiment with artificial neural networks and adjust weights to get the desired outcomes. The purpose of this project is to use an artificial neural network and a complementary metal oxide semiconductor (CMOS) to develop a decoder gate.We constantly keep in mind that the goal should be to construct a COMS circuit at the lowest feasible cost.An artificial neural network with multiple layers was used to create the circuit in this work. In our study, we treat negative values as inverters and neurons as transistors, and we modify the value using weights.Additionally, we are creating Verilog-HdL code that will make it simple to use the Decoder to experiment with artificial neural networks and adjust weights to get the desired outcomes. The purpose of this research is to use an artificial neural network and a complementary metal oxide semiconductor (CMOS) to produce a multiplexer gate.We always keep in mind that the lowest feasible cost should be the goal while designing any COMS circuit.This work used a multilayer artificial neural network to design the circuit. In our work, we treat neurons as transistors and negative values as inverters, and we use weights to adjust the value.In order to effortlessly apply the Multiplexer for experimenting with an artificial neural network and set weights to obtain the desired results, we are also building Verilog-HdL code.Keywords: Multiplexer, CMOS,ANN,Verilog-HDL. Keywords: Multiplexer De-multiplexer, Decoder, Encoder Verilog-HDL, CMOS, ANN, Verilog HDL

Image classification system for COVID-19 patient X-Ray results using median filter with deep learning method

An image is a representation of an object that is rewritten on a medium with specific values (intensity) and has x and y coordinates. X-ray images are one type of medical image that can be used to detect and study a disease. However, X-ray images can sometimes appear blurry, making interpretation a bit challenging. Furthermore, there is variation in X-ray attenuation between normal tissue and tissue affected by a disease. By using the median filter and implementing deep learning as a method for classifying images based on feature extraction and weights in artificial neural networks. The stages carried out include preprocessing, with training and testing, feature extraction using layers of artificial neural networks that have undergone the filtering process. Then, a custom classifier layer is created to train the classification of COVID and normal classes using data from bottleneck.npy. The results of training and testing yielded an average accuracy of 95%, while the F1 Score evaluation result was 0.9."

On the effects of biased quantum random numbers on the initialization of artificial neural networks

Recent advances in practical quantum computing have led to a variety of cloud-based quantum computing platforms that allow researchers to evaluate their algorithms on noisy intermediate-scale quantum devices. A common property of quantum computers is that they can exhibit instances of true randomness as opposed to pseudo-randomness obtained from classical systems. Investigating the effects of such true quantum randomness in the context of machine learning is appealing, and recent results vaguely suggest that benefits can indeed be achieved from the use of quantum random numbers. To shed some more light on this topic, we empirically study the effects of hardware-biased quantum random numbers on the initialization of artificial neural network weights in numerical experiments. We find no statistically significant difference in comparison with unbiased quantum random numbers as well as biased and unbiased random numbers from a classical pseudo-random number generator. The quantum random numbers for our experiments are obtained from real quantum hardware.

A hybrid GA-ANN and correlation approach to developing a laminar burning velocity prediction model for isooctane/blends-air mixtures

Surrogate fuels offer a cost-effective way to predict the combustion properties of transportation fuels like diesel, gasoline, kerosene, etc. Iso-octane (2,2,4-trimethylpentane) is a key gasoline reference and surrogate component. Researchers explore alternative fuels and their combustion characteristics to enhance efficiency and reduce emissions. The rising interest lies in the blending of isooctane with various alternative fuels, aiming for cleaner and more efficient combustion. In this current study, a machine learning method called Feed-Forward Artificial Neural Network (FFANN) with back-propagation (BP) was employed to forecast the laminar burning velocity (LBV) of isooctane/blends – air mixtures. A total of eleven blends including ammonia, hydrogen, methane, methanol, ethanol, butanol, n-heptane, 2-methyl furan (2-MF), 2,5-dimethylfuran (2,5-DMF), 2-methyl tetrahydrofuran (2-MTHF), and syngas were examined. The artificial neural network (ANN) model was created using a dataset consisting of 2234 data points gathered from the past experimental literature since 1983. To enhance the ANN's predictive capability, a combination of the random search CV technique and selective testing approach was utilized for optimizing ANN hyperparameters, while the genetic algorithm (GA) was deployed to optimize the ANN's weight values. The development of the ANN model was carried out within the Python software environment, utilizing the Keras application programming interface. The constructed GA-ANN model was compared to a variety of other machine learning (ML) models developed within this study, including generalized linear regression (GLR), support vector machine (SVM), random forest (RF), artificial neural network (ANN), and XGBoost regression. When evaluated on the testing set, which constitutes 15% of the complete dataset, the GA-ANN model demonstrated superior performance compared to all other ML models utilized in this research, achieving an impressive prediction accuracy with the coefficient of determination (R2) of 0.9910, root mean square error (RMSE) of 0.8231, and a mean absolute error (MAE) of 0.643. Additionally, a common LBV correlation for all isooctane/blend-air mixtures was created using the extended Gulder’s LBV formulation with the input parameters including pressure, temperature, equivalence ratio, and molar fraction of blend. This correlation results showed very minimum deviation from the experimental with an RMSE value of less than 0.38541.

Obtaining an accurate prediction model for viscosity of a new nano-lubricant containing multi-walled carbon nanotube-titanium dioxide nanoparticles with oil SAE50

This study aims to investigate the viscosity behavior of multi-walled carbon nanotube (MWCNT) - titanium dioxide (TiO2) (40−60) - SAE50 oil nanofluid using an Artificial Neural Network (ANN) modeling approach. The main objective is to develop a highly accurate predictive model for viscosity by considering three input parameters: temperature, solid volume fraction (SVF), and shear rate (SR). Rheological measurements provide experimental data used to train and validate the ANN model. The ANN model's architecture, activation functions, and training algorithms are carefully chosen. Data are divided to three subsets including train, validation and test. ANN is trained using trainlm algorithm for 50 times to vanish the effect of random nature of ANN weight initialization. The trained ANN model is then utilized to predict the viscosity of the nanofluid under varying conditions. The results demonstrate the efficacy of the proposed ANN model in capturing the complex relationship between viscosity and the input parameters, providing accurate viscosity predictions for the MWCNT-TiO2-oil SAE50 nanofluid. Furthermore, the influence of temperature, SVF, and SR on viscosity is analyzed, offering valuable insights into the flow behavior of the nanofluid. According to the obtained results, the developed ANN model presents a reliable and efficient approach to estimate the viscosity of the MWCNT-TiO2-SAE50 oil nanofluid, eliminating the need for costly and extensive experimental measurements within the analyzed range. ANN could model the nanofluid viscosity with R2 = 0.9998 and MSE= 0.000189 that is quite acceptable. Also, the experimental data revealed that for the investigated nanofluid, temperature and shear rate have impressive effect on the viscosity (changing viscosity more than 100% for the analyzed margin), on the other hand, the nanoparticle volume fraction effect is much lower, to be more precise, increasing the nanoparticle percentage will increase the viscosity mean value around 30%.

Machine learning intelligent based hydromagnetic thermal transport under Soret and Dufour effects in convergent/divergent channels: a hybrid evolutionary numerical algorithm

In this research, we analyze the complex dynamics of hydro-magnetic flow and heat transport under Sorent and Dofour effects within wedge-shaped converging and diverging channels emphasizing its critical role in conventional system design, high-performance thermal equipment. We utilized artificial neural networks (ANNs) to investigation the dynamics of the problem. Our study centers on unraveling the intricacies of energy transport and entropy production arising from the pressure-driven flow of a non-Newtonian fluid within both convergent and divergent channel. The weights of ANN based fitness function ranging from − 10 to 10. To optimize the weights and biases of artificial neural networks (ANNs), employ a hybridization of advanced evolutionary optimization algorithms, specifically the artificial bee colony (ABC) optimization integrated with neural network algorithms (NNA). This approach allows us to identify and fine-tune the optimal weights within the neural network, enabling accurate prediction. We compare our results against the established different analytical and numerical methods to assess the effectiveness of our approach. The methodology undergoes a rigorous evaluation, encompassing multiple independent runs to ensure the robustness and reliability of our findings. Additionally, we conduct a comprehensive analysis that includes metrics such as mean squared error, minimum values, maximum values, average values, and standard deviation over these multiple independent runs. The minimum fitness function value is 1.32 × 10−8 computed across these multiple runs. The absolute error, between the HAM and machine learning approach addressed ranging from 3.55 × 10−7 to 1.90 × 10−8. This multifaceted evaluation ensures a thorough understanding of the performance and variability of our proposed approach, ultimately contributing to our understanding of entropy management in non-uniform channel flows, with valuable implications for diverse engineering applications.

Adaptive Neuroevolution With Genetic Operator Control and Two-Way Complexity Variation

Topology and weight evolving artificial neural network (TWEANN) algorithms optimize the structure and weights of artificial neural networks (ANNs) simultaneously. The resulting networks are typically used as policy models for solving control and reinforcement learning (RL) type problems. This paper presents a neuroevolution algorithm that aims to address the typical stagnation and sluggish convergence issues present in other neuroevolution algorithms. These issues are often caused by inadequacies in population diversity preservation, exploration/exploitation balance, and search flexibility. This new algorithm, called the Adaptive Genomic Evolution of Neural-Network Topologies (AGENT), builds on the neuroevolution of augmenting topologies (NEAT) concept. Novel mechanisms for adapting the selection and mutation operations are proposed to favorably control population diversity and exploration/exploitation balance. The former is founded on a fundamentally new way of quantifying diversity by taking a graph-theoretic perspective of the population of genomes and inter-genomic differences. Further advancements to the NEAT paradigm occur through the incorporation of variable neuronal properties and new mutation operations that uniquely allow both the growth and pruning of ANN topologies during evolution. Numerical experiments with benchmark control problems adopted from the OpenAI Gym illustrate the competitive performance of AGENT against standard RL methods and adaptive HyperNEAT, and superiority over the original NEAT algorithm. Further parametric analysis provides key insights into the impact of the new features in AGENT. This is followed by evaluation on an unmanned aerial vehicle collision avoidance problem where maneuver planning models are learnt by AGENT with 33% reward improvement over 15 generations.

On Neuroevolution of Multi-Input Compositional Pattern Producing Networks: A Case of Entertainment Computing, Edge Devices, and Smart Cities

This work presents a novel approach by utilizing Heterogeneous Activation Neural Networks (HA-NNs) to evolve the weights of Artificial Neural Networks (ANNs) for reinforcement learning in console and arcade computer games like Atari's Breakout and Sonic the Hedgehog. It is the first study to explore the potential of HA-NNs as potent ANNs in solving gaming-related reinforcement learning problems. Additionally, the proposed solution optimizes data transmission over networks for edge devices, marking a novel application of HA-NNs. The study achieved outstanding results, outperforming recent works in benchmark environments like CartPole-v1, Lunar Lander Continuous, and MountainCar-Continuous, with HA-NNs and ANNs evolved using the Neuroevolution of Augmenting Topologies (NEAT) algorithm. Notably, the key advancements include exceptional scores of 500 in CartPole-v1 and 98.2 in Mountain Car Continuous, demonstrating the efficacy of HA-NNs in reinforcement learning tasks. Beyond gaming, the research addresses the challenge of efficient data communication between edge devices, which has the potential to enhance performance in smart cities while reducing the load on edge devices and supporting seamless entertainment experiences with minimal commuting. This work pioneers the application of HA-NNs in reinforcement learning for computer games and introduces a novel approach for optimizing edge device communication, promising significant advancements in the fields of AI, neural networks, and smart city technologies.

The Prediction of Cu(II) Adsorption Capacity of Modified Pomelo Peels Using the PSO-ANN Model.

It is very well known that traditional artificial neural networks (ANNs) are prone to falling into local extremes when optimizing model parameters. Herein, to enhance the prediction performance of Cu(II) adsorption capacity, a particle swarm optimized artificial neural network (PSO-ANN) model was developed. Prior to predicting the Cu(II) adsorption capacity of modified pomelo peels (MPP), experimental data collected by our research group were used to build a consistent database. Then, a PSO-ANN model was established to enhance the model performance by optimizing the ANN's weights and biases. Finally, the performances of the developed ANN and PSO-ANN models were deeply evaluated. The results of this investigation revealed that the proposed hybrid method did increase both the generalization ability and the accuracy of the predicted data of the Cu(II) adsorption capacity of MPPs when compared to the conventional ANN model. This PSO-ANN model thus offers an alternative methodology for optimizing the adsorption capacity prediction of heavy metals using agricultural waste biosorbents.

A Low-Power ECG Processor ASIC Based on an Artificial Neural Network for Arrhythmia Detection

The early detection of arrhythmia can effectively reduce the risk of serious heart diseases and save time for treatment. Many healthcare devices have been widely used for electrocardiogram (ECG) monitoring. However, most of them can only complete simple two-classes detection and have unacceptable hardware overhead and energy consumption. For achieving accurate and low-power arrhythmia detection, a novel ECG processor application specific integrated circuit (ASIC) is proposed in this paper, which can perform the prediction of five types of cardiac arrhythmias and heart rate monitoring. To realize hardware-efficient R-peak detection, an ECG pre-processing engine based on a first derivative and moving average comparison method is proposed. Efficient arrhythmia detection is realized by the proposed low-power classification engine, which is based on a carefully designed lightweight artificial neural network (ANN) with good prediction accuracy. The hardware reuse strategy is used to implement the hardware logic of ANN, where computations are executed by only one processing unit (PU), which is controlled by a flexible finite state machine (FSM). Also, the weights of ANN are configurable to facilitate model updates. We validate the functionality of the design using real-world ECG data. The proposed ECG processor is implemented using 55 nm CMOS technology, occupying an area of 0.33 mm2. This design consumes 12.88 μW at a 100 kHz clock frequency, achieving a classification accuracy of 96.69%. The comparison results with previous work indicate that our design has advantages in detection performance and power consumption, providing a good solution for low-power and low-cost ECG monitoring.

A novel neural network model with distributed evolutionary approach for big data classification

The considerable improvement of technology produced for various applications has resulted in a growth in data sizes, such as healthcare data, which is renowned for having a large number of variables and data samples. Artificial neural networks (ANN) have demonstrated adaptability and effectiveness in classification, regression, and function approximation tasks. ANN is used extensively in function approximation, prediction, and classification. Irrespective of the task, ANN learns from the data by adjusting the edge weights to minimize the error between the actual and predicted values. Back Propagation is the most frequent learning technique that is used to learn the weights of ANN. However, this approach is prone to the problem of sluggish convergence, which is especially problematic in the case of Big Data. In this paper, we propose a Distributed Genetic Algorithm based ANN Learning Algorithm for addressing challenges associated with ANN learning for Big data. Genetic Algorithm is one of the well-utilized bio-inspired combinatorial optimization methods. Also, it is possible to parallelize it at multiple stages, and this may be done in an extremely effective manner for the distributed learning process. The proposed model is tested with various datasets to evaluate its realizability and efficiency. The results obtained from the experiments show that after a specific volume of data, the proposed learning method outperformed the traditional methods in terms of convergence time and accuracy. The proposed model outperformed the traditional model by almost 80% improvement in computational time.

Identification and control of Maglev system using fractional and integer order PID controller

System identification techniques have proved to be the most effective methodologies for the modeling highly non-linear and system. For the purpose of real-time parameter estimation of a Maglev system, a Teaching Learning Based Optimization (TLBO) for updating the weights of Functional Link Artificial Neural Network (FLANN) model is proposed and implemented in this research. Moreover, we proposed a one & two-Degree of Freedom (one-DOF & two-DOF) Fractional Order PID (FOPID) controller, where the parameters are optimized by using the Teaching Learning Based Optimization (TLBO) and the recently proposed Black Widow Optimization (BWO) algorithm. To investigate the robustness of the proposed controller, a pulse signal disturbance is added at equal intervals of the output of the identified model of the Maglev system. It is found that the suggested two-DOF FOPID controller with TLBO performs better than its counterpart in terms of both in time domain specifications (i.e., maximum overshoot = 1.2648%, settling time = 1.3884 sec and rise time = 0.8685 sec) and in robustness analysis (i.e., system is sufficiently robust, because the infinity norms of the sensitivity and the complementary sensitivity functions of the system are less than two). The TLBO algorithm has performed better for both identification and optimization of controller parameter due to very less number of algorithmic parameter is as compared to other algorithm.

Modeling and optimization of CO2 mass transfer flux into Pz-KOH-CO2 system using RSM and ANN

In this research, artificial neural networks (ANN) and response surface methodology (RSM) were applied for modeling and optimization of carbon dioxide (CO2) absorption using KOH-Pz-CO2 system. In the RSM approach, the central composite design (CCD) describes the performance condition in accordance with the model using the least-squares technique. The experimental data was placed in second-order equations applying multivariate regressions and appraised applying analysis of variance (ANOVA). The p-value for all dependent variables was obtained to be less than 0.0001, indicating that all models were significant. Furthermore, the experimental values obtained for the mass transfer flux satisfactorily matched the model values. The R2 and Adj-R2 models are 0.9822 and 0.9795, respectively, which, it means that 98.22% of the variations for the NCO2 is explained by the independent variables. Since the RSM does not create any details about the quality of the solution acquired, the ANN method was applied as the global substitute model in optimization problems. The ANNs are versatile utensils that can be utilized to model and anticipate different non-linear and involved processes. This article addresses the validation and improvement of an ANN model and describes the most frequently applied experimental plans, about their restrictions and generic usages. Under different process conditions, the developed ANN weight matrix could successfully forecast the behavior of the CO2 absorption process. In addition, this study provides methods to specify the accuracy and importance of model fitting for both methodologies explained herein. The MSE values for the best integrated MLP and RBF models for the mass transfer flux were 0.00019 and 0.00048 in 100 epochs, respectively.

Gradient free stochastic training of ANNs, with local approximation in partitions

We present a numerical scheme for computation of Artificial Neural Networks (ANN) weights, which stems from the Universal Approximation Theorem, avoiding costly iterations. The proposed algorithm adheres to the underlying theory, is highly fast, and results in remarkably low errors when applied to regression and classification problems of complex data sets with {textbf{x}} in {mathbb {R}}^{n} (e.g. Griewank, Gomez-Levy, Shekel, and Polynomial functions) with random noise addition (i.e. Uniform, Normal, Generalized Pareto, Log-Normal, and a mixture of Log-Normal, Exponential, and Frechet), as well as the database for handwritten digits recognition MNIST (Modified National Institute of Standards and Technology) with 7times 10^4 images. The same mathematical formulation was found capable of approximating highly nonlinear functions in multiple dimensions, with low errors (e.g. 10^{-10}) for the test set of the unknown functions, their higher-order partial derivatives, as well as numerically solving Partial Differential Equations, such as those appearing in Physics, Engineering, Environmental Sciences, etc. The method is based on the calculation of the weights of each neuron in small neighbourhoods of the data. Accordingly, optimization of hyperparameters is not necessary, as the number of neurons stems directly from the dimensionality of the data, further improving the algorithmic speed. Under this setting, overfitting is inherently avoided, and the results are interpretable and reproducible. The complexity of the proposed algorithm is of class P with {mathcal {O}}(mNni_{cl} + Nmn^2+Nn^3 + mN^2+N^3) computing time, with respect to the observations m, features n, and Neurons N, contrary to the NP-Complete class of standard algorithms for ANN training. The performance of the method is high, irrespective of the size of the data set, and the test set errors are similar or smaller than the training errors, indicating the generalization efficiency of the algorithm. A supplementary computer code in Julia and Python Languages is provided, which can be used to reproduce the validation examples, and/or apply the algorithm to other data sets.

Artificial intelligence techniques for encrypt images based on the chaotic system implemented on field-programmable gate array

<span lang="EN-US">Image encryption is an important issue in protecting the content of images and in the area of information security. This article proposes a novel method for image encryption and decryption using the structure of the artificial neural network (ANN)-based chua chaotic system (CCS). This structure was efficiently designed on a field-programmable gate array (FPGA) chip utilizing the xilinx system generator (XSG) tool with the IEEE-754-1985 32-bit floating-point number format. For ANN-based CCS design, a multilayer feed forward neural network (FFNN) structure with three inputs and three outputs was created. This structure consists of one hidden layer with four neurons, each of which has a Tangent Sigmoid activation function. The training of ANN-based CCS yielded a 3.602e-13 mean square error (MSE) value. After successfully training the ANN-based CCS, the design was carried out on FPGA, utilizing the ANN structure's bias and weight values as a reference. The xilinx vivado (2017.4) design suite was used to synthesis and test the ANN-based CCS on the FPGA. The histogram, correlation coefficient, and entropy are used to perform security analysis on various images. Finally, FPGA hardware co-simulation using a Xilinx Artix7 xc7a100t-1csg324 chip was utilized to verify that the encryption and decryption of the images were successful.</span>