Implementation Of Artificial Neural Networks Research Articles

Implementation of artificial neural network in a building benefits from radiant floor heating /cooling enhanced by phase change materials

In this study, the numerical analysis of the radiant floor system was investigated for a building in the presence of PCM inside the external walls as well as the roof at a thickness of 2 cm. By injecting cold/warm fluid into the radiant tubes inside the roof, the cooling/heating requirements were met. Several PCMs with identical thermal properties (except melting point) were selected and based on numerical analysis, the energy utilization in the heating/cooling sections was evaluated by comparison with the simple building (without PCM). Four main variables were defined for the neural network, and energy consumption was trended for two climate zones, Shenyang (41.7922°N, 123.4328°E), and Zhengzhou (34.7578°N, 113.6486° E). For each region, the PCM with the best phase transition was selected and it was realized that for the first region, energy consumption was diminished by 12.6% and for the second region by 15.9%. According to the temperature conditions and radiation intensity in the environment, the ANN could forecast annual energy utilization with an error of less than 6%.

From an Engineer's Perspective: Evaluation of the Invested Effort and the Resulting Benefits of Artificial Neural Networks in Proton Exchange Membrane Water Electrolysis Development

Machine learning approaches have become increasingly important in engineering research and development in recent years, and the first steps in this field can also be found at proton exchange membrane water electrolysis (PEMWE) development [1]. Well-known conventional simulation models for engineers are usually based on physical and empirical equations, with different approaches depending on the desired model depth. However, especially for the simulation of complex processes, model formulation and parametrization with sufficient accuracy are often challenging, with these types of models reaching their limits. In contrast, data-driven models use machine learning approaches to reproduce the desired processes based on large data sets and do not require physical or empirical model formulations [2]. For decision making on the deployment of machine learning methods, the effort required for familiarization, conceptual design, and implementation is the crucial aspect to answering the question of beneficial gain. Our conference contribution aims to answer these questions regarding invested effort and resulting benefit through a self-experiment using the example of artificial neural networks (ANN) in PEMWE research. It is intended to guide engineers interested in machine learning methods without prior knowledge, addressing their questions and ultimately assisting them in decision-making.Our proposed research is shown in Figure 1 and consists of conceptualizing and programming ANN for two application cases, one for high-frequency resistance (HFR) regression and the other for PEMWE single-cell aging prediction. Here, the processes of familiarization time, conceptualization, and implementation of ANN are transparently presented from an engineer's perspective. The modeling results are evaluated in terms of their utility to the user. The data sets for training and validating the ANN to calculate HFR are generated using a physics-based simulation model. For aging prediction, the datasets are based on experimental data obtained from accelerated stress tests (AST) of a PEMWE single cell.Our contribution finalizes with an assessment of the profit growth of engineering applications for pioneers in the field of machine learning methods.Figure 1. Scheme for the evaluation of PEMWE single cells by artificial neural networks from an engineering perspective.

One-Phototransistor-One-Memristor Array with High-Linearity Light-Tunable Weight for Optic Neuromorphic Computing.

The recent advances in optic neuromorphic devices have led to a subsequent rise in use for construction of energy-efficient artificial-vision systems. The widespread use can be attributed to their ability to capture, store, and process visual information from the environment. The primary limitations of existing optic neuromorphic devices include nonlinear weight updates, cross-talk issues, and silicon process incompatibility. In this study, a highlylinear, light-tunable, cross-talk-free, and silicon-compatible one-phototransistor-one-memristor (1PT1R) optic memristor is experimentally demonstrated for the implementation of an optic artificial neural network (OANN). For optic image recognition in the experiment, an OANN is constructed using a 16×3 1PT1R memristor array, and it is trained on an online platform. The model yields an accuracy of 99.3% after only ten training epochs. The 1PT1R memristor, which shows good performance, demonstrates its ability as an excellent hardware solution for highly efficient optic neuromorphic and edge computing.

On the Application of Artificial Neural Network for Classification of Incipient Faults in Dissolved Gas Analysis of Power Transformers

Oil-submerged transformer is one of the inherent instruments in the South African power system. Transformer malfunction or impairment may interpose the operation of the electric power distribution and transmission system, coupled with liability for high overhaul costs. Hence, recognition of inchoate faults in an oil-submerged transformer is indispensable and it has turned into an intriguing subject of interest by utility owners and transformer manufacturers. This work proposes a hybrid implementation of a multi-layer artificial neural network (MLANN) and IEC 60599:2022 gas ratio method in identifying inchoate faults in mineral oil-based submerged transformers by employing the dissolved gas analysis (DGA) method. DGA is a staunch practice to discover inchoate faults as it furnishes comprehensive information in examining the transformer state. In current work, MLANN was established to pigeonhole seven fault types of transformer states predicated on the three International Electrotechnical Commission (IEC) combustible gas ratios. The designs enmesh the development of numerous MLANN algorithms and picking networks with the optimum performance. The gas ratios are in accordance with the IEC 60599:2022 standard whilst an empirical databank comprised of 100 datasets was used in the training and testing activities. The designated MLANN design produces an overall correlation coefficient of 0.998 in the categorization of transformer state with reference to the combustible gas produced.

Synthetic neuromorphic computing in living cells

Computational properties of neuronal networks have been applied to computing systems using simplified models comprising repeated connected nodes, e.g., perceptrons, with decision-making capabilities and flexible weighted links. Analogously to their revolutionary impact on computing, neuro-inspired models can transform synthetic gene circuit design in a manner that is reliable, efficient in resource utilization, and readily reconfigurable for different tasks. To this end, we introduce the perceptgene, a perceptron that computes in the logarithmic domain, which enables efficient implementation of artificial neural networks in Escherichia coli cells. We successfully modify perceptgene parameters to create devices that encode a minimum, maximum, and average of analog inputs. With these devices, we create multi-layer perceptgene circuits that compute a soft majority function, perform an analog-to-digital conversion, and implement a ternary switch. We also create a programmable perceptgene circuit whose computation can be modified from OR to AND logic using small molecule induction. Finally, we show that our approach enables circuit optimization via artificial intelligence algorithms.

Spiking neural networks based on two-dimensional materials

The development of artificial neural networks using memristors is gaining a lot of interest among technological companies because it can reduce the computing time and energy consumption. There is still no memristor, made of any material, capable to provide the ideal figures-of-merit required for the implementation of artificial neural networks, meaning that more research is required. Here we present the use of multilayer hexagonal boron nitride based memristors to implement spiking neural networks for image classification. Our study indicates that the recognition accuracy of the network is high, and that can be resilient to device variability if the number of neurons employed is large enough. There are very few studies that present the use of a two-dimensional material for the implementation of synapses of different features; in our case, in addition to a study of the synaptic characteristics of our memristive devices, we deal with complete spiking neural network training and inference processes.

Implementation of Backpropagation ANN in Predicting Long Bean Crop Production in Sumatra Island Province

The production of long bean vegetable crops in Indonesia is very high, this is because this plant is easy to cultivate. Predicting the production of long bean vegetable crops on the island of Sumatra, where the data source comes from BPS (Central Bureau of Statistics). In predicting the use of ANN (Artificial Neural Networks) and the method used in this study is the backpropagation algorithm, this method will be used to predict or predict the production of long bean vegetable crops on the island of Sumatra. The results have been obtained using 4 models, namely the 6-5-1, 6-10-1, 6-15-1, and 6-20-1 models. Among the 4 existing models, the 6-5-1 model has the more accurate accuracy or the lowest error value with an MSE of 0.00711838.

Artificial Neural Network (ANN) Implementation with Conjugate Gradient Algorithm to Predict Sumatran Melinjo Plant Production

Melinjo is an annual plant with open seeds. Tree-shaped and has two houses called dioecious or there are males and females. Melinjo is often found in dry and tropical areas. Indonesia can be one that produces melinjo as a trade product in large quantities. Melinjo is collected and shipped natural products after 5-6 long time after sowing of seeds. In West Sumatra, it is detailed that each year produces 20,000 to 25,000 natural melinjo products and the seed generation reaches 80 to 100 kg per tree per year. Therefore, it is important to know every need for melinjo by anticipating the number of generations of Melinjo using a Manufacturing Artificial Neural System with Backpropagation strategy. With the neural structure made, it will be easier to carry out this investigation. Where the machine learning method can help to find the best performance value and value from the simple data studied. The Matlab2011b application has a feature that helps to calculate the best performance and value with the help of the Conjugate Gradient algorithm. After testing using 5 samples, namely: 4-10-1, 4-15-1, 4-20-1,4-25-1, 4-30-1. Of the five tests, the best results are on data 4-15-1 with the MSE/Performance value of 0.011154591. 4-15-1, 4-20-1,4-25-1, 4-30-1. Of the five tests, the best results are on data 4-15-1 with the MSE/Performance value of 0.011154591. 4-15-1, 4-20-1,4-25-1, 4-30-1. Of the five tests, the best results are on data 4- 15-1 with the MSE/Performance value of 0.011154591.

Deep Learning Concepts and Applications for Synthetic Biology.

Synthetic biology has a natural synergy with deep learning. It can be used to generate large data sets to train models, for example by using DNA synthesis, and deep learning models can be used to inform design, such as by generating novel parts or suggesting optimal experiments to conduct. Recently, research at the interface of engineering biology and deep learning has highlighted this potential through successes including the design of novel biological parts, protein structure prediction, automated analysis of microscopy data, optimal experimental design, and biomolecular implementations of artificial neural networks. In this review, we present an overview of synthetic biology-relevant classes of data and deep learning architectures. We also highlight emerging studies in synthetic biology that capitalize on deep learning to enable novel understanding and design, and discuss challenges and future opportunities in this space.

Trends of Microwave Devices Design Based on Artificial Neural Networks: A Review

The usage of techniques of the artificial neural networks (ANNs) in the field of microwave devices has recently increased. The advantages of ANNs in comparison with traditional full-wave methods are that the prediction speed when the traditional time-consuming iterative calculations are not required and also the complex mathematical model of the microwave device is no longer needed. Therefore, the design of microwave device could be repeated many times in real time. However, methods of artificial neural networks still lag behind traditional full-wave methods in terms of accuracy. The prediction accuracy depends on the structure of the selected neural network and also on the obtained dataset for the training of the network. Therefore, the paper presents a systematic review of the implementation of ANNs in the field of the design and analysis of microwave devices. The guidelines for the systematic literature review and the systematic mapping research procedure, as well as the Preferred Report Items for Systematic Reviews and Meta-Analysis statements (PRISMA) are used to conduct literature search and report the results. The goal of the paper is to summarize the application areas of usage of ANNs in the field of microwave devices, the type and structure of the used artificial neural networks, the type and size of the dataset, the interpolation and the augmentation of the training dataset, the training algorithm and training errors and also to discuss the future perspectives of the usage of ANNs in the field of microwave devices.

Frequency-Dependent Synapse Weight Tuning in 1S1R with a Short-Term Plasticity TiOx-Based Exponential Selector.

Short-term plasticity (STP) is an important synaptic characteristic in the hardware implementation of artificial neural networks (ANN), as it enables the temporal information processing (TIP) capability. However, the STP feature is rather challenging to reproduce from a single nonvolatile resistive random-access memory (RRAM) element, as it requires a certain degree of volatility. In this work, a Pt/TiOx/Pt exponential selector is introduced not only to suppress the sneak current but also to enable the TIP feature in a one selector-one RRAM (1S1R) synaptic device. Our measurements reveal that the exponential selector exhibits the STP characteristic, while a Pt/HfOx/Ti RRAM enables the long-term memory capability of the synapse. Thereafter, we experimentally demonstrated pulse frequency-dependent multilevel switching in the 1S1R device, exhibiting the TIP capability of the developed 1S1R synapse. The observed STP of the selector is strongly influenced by the bottom metal-oxide interface, in which Ar plasma treatment on the bottom Pt electrode resulted in the annihilation of the STP feature in the selector. A mechanism is thus proposed to explain the observed STP, using the local electric field enhancement induced at the metal-oxide interface coupled with the drift-diffusion model of mobile O2- and Ti3+ ions. This work therefore provides a reliable means of producing the STP feature in a 1S1R device, which demonstrates the TIP capability sought after in hardware-based ANN.

Improving FPGA Based Impedance Spectroscopy Measurement Equipment by Means of HLS Described Neural Networks to Apply Edge AI

The artificial intelligence (AI) application in instruments such as impedance spectroscopy highlights the difficulty to choose an electronic technology that correctly solves the basic performance problems, adaptation to the context, flexibility, precision, autonomy, and speed of design. Present work demonstrates that FPGAs, in conjunction with an optimized high-level synthesis (HLS), allow us to have an efficient connection between the signals sensed by the instrument and the artificial neural network-based AI computing block that will analyze them. State-of-the-art comparisons and experimental results also demonstrate that our designed and developed architectures offer the best compromise between performance, efficiency, and system costs in terms of artificial neural networks implementation. In the present work, computational efficiency above 21 Mps/DSP and power efficiency below 1.24 mW/Mps are achieved. It is important to remark that these results are more relevant because the system can be implemented on a low-cost FPGA.

Hardware implementation of neural network-based engine model using FPGA

This paper implements an artificial neural network (ANN)-based engine model using the Field Programmable Gate Array (FPGA). The developed (ANN)-based engine model will be used to estimate the engine gas emissions to mitigate the harmful effects of these emissions on human health. Getting reliable and robust FPGA-based ANNs implementations depends on the optimal choice of activation function that will provide minimal area occupation on FPGA. This study introduces, implements, and investigates FPGA-based ANN-based engine models using five different activation functions. These implemented engine models were described using MATLAB/Simulink and hardware description language coder and carried out by Spartan -3E-500.CP132 FPGA platform from Xilinx. The performance of the implemented engine models was investigated in terms of area-efficient implementation and the regression values (R) to build a robust ANN-based engine model.

A survey on neural networks for (cyber-) security and (cyber-) security of neural networks

The goal of this systematic and broad survey is to present and discuss the main challenges that are posed by the implementation of Artificial Intelligence and Machine Learning in the form of Artificial Neural Networks in Cybersecurity, specifically in Intrusion Detection Systems. Based on the results of the state-of-the-art analysis with a number of bibliographic methods, as well as their own implementations, the authors provide a survey of the answers to the posed problems as well as effective, experimentally-found solutions to those key issues. The issues include hyperparameter tuning, dataset balancing, increasing the effectiveness of an ANN, securing the networks from adversarial attacks, and a range of non-technical challenges of applying ANNs for IDS, such as societal, ethical and legal dilemmas, and the question of explainability. Thus, it is a systematic review and a summary of the body of knowledge amassed around implementations of Artificial Neural Networks in Network Intrusion Detection, guided by an actual, real-world implementation.

P‐23: A Monolithically Integrated Artificial Compound Eye for Proximity Pattern Recognition

A biomimetic artificial compound eye is constructed as an array of pixels each consisting of in‐pixel circuits made of several metal‐oxide thin‐film transistors, a hydrogenated amorphous silicon photodiode and a structured optical path. An in‐pixel amplifier is incorporated to enhance the signal‐to‐noise ratio. The peak fabrication temperature of 300 �C is compatible with flexible substrates. Using an artificial neural network implementation, the compound eye has been trained for proximity pattern recognition at a separation of 12 mm and achieving an inference accuracy of 95%. Sensing at a larger separation is being implemented by re‐designing the optical path, leading to potential recognition of dynamic gestures.

Implementation of Artificial Neural Network to Predict Diabetes with High-Quality Health System.

Patients with diabetes who are closely monitored have a higher overall quality of life than those who are not. Costs associated with healthcare can be decreased by utilising the Internet of Things (IoT), thanks to technological advancements. To satisfy the expectations of e-health applications, it is required for the development of the intelligent systems as well as increases the number of applications that are connected to the network. As a result, in order to achieve these goals, the cellular network should be capable of supporting intelligent healthcare applications that require high energy efficiency. In this paper, we model a neural network-based ensemble voting classifier to predict accurately the diabetes in the patients via online monitoring. The study consists of Internet of Things (IoT) devices to monitor the instances of the patients. While monitoring, the data are transferred from IoT devices to smartphones and then to the cloud, where the process of classification takes place. The simulation is conducted on the collected samples using the python tool. The results of the simulation show that the proposed method achieves a higher accuracy rate, higher precision, recall, and f-measure than existing state-of-art ensemble models.

Improving HfO2-Based Resistive Switching Devices by Inserting a TaOx Thin Film via Engineered In Situ Oxidation

Resistive switching (RS) devices with binary and analogue operation are expected to play a key role in the hardware implementation of artificial neural networks. However, state of the art RS devices based on binary oxides (e.g., HfO2) still do not exhibit enough competitive performance. In particular, variability and yield still need to be improved to fit industrial requirements. In this study, we fabricate RS devices based on a TaOx/HfO2 bilayer stack, using a novel methodology that consists of the in situ oxidation of a Ta film inside the atomic layer deposition (ALD) chamber in which the HfO2 film is deposited. By means of X-ray reflectivity (XRR) and time-of-flight secondary ion mass spectrometry (ToF-SIMS), we realized that the TaOx film shows a substoichiometric structure, and that the TaOx/HfO2 bilayer stack holds a well-layered structure. An exhaustive electrical characterization of the TaOx/HfO2-based RS devices shows improved switching performance compared to the single-layer HfO2 counterparts. The main advantages are higher forming yield, self-compliant switching, lower switching variability, enhanced reliability, and better synaptic plasticity.

The Implementation of Artificial Neural Network (ANN) on Offline Cursive Handwriting Image Recognition

Identifying a writing is an easy thing to do for human, but this does not apply to computers, in particular if it is handwriting. Handwriting recognition, especially cursive handwriting is a research in the area of image processing and pattern matching that is challenging to complete, following the different characteristics of each person's cursive handwriting style. In this study, the use of the ANN model will be implemented in performing offline handwriting image recognition. The cursive handwriting image that has been obtained is then preprocessed and segmented using bounding box rectangle and contour techniques. Evaluation of system performance using global performance metrics in this study resulted in a percentage of 93% where the bounding box and contour succeeded in determining the segmentation point correctly, so that the ANN model worked optimally.

Numerical study of the effects of twisted-tape inserts on heat transfer parameters and pressure drop across a tube carrying Graphene Oxide nanofluid: An optimization by implementation of Artificial Neural Network and Genetic Algorithm

A numerical study is undertaken to investigate the effects of twisted-tape inserts on the heat transfer and pressure drop across a horizontal tube carrying water-based Graphene-Oxide nanofluid. The flow of the nanofluid is considered for dissimilar volume fraction and Re. Twisted tapes of various aspect ratios are placed inside the tube. The results show that, at high volume fractions and Re levels, solid nanoparticles improve the heat transfer provided that the twist ratio is not very small. The highest heat transfer is achieved at twist ratio of 2.34 and nanoparticle volume fraction of 4% regardless of Re, and the highest PEC is corresponded to the twist ratio of 2.34, Re of 5000, and 4 % volume fraction. Then, the Artificial Neural Network is used for the estimation of the Nu, pressure drop, and PEC. Based on the results, Nu rises by increasing Re and the volume fraction, while the friction factor rises by reducing the Re and the twist pitch. Finally, an optimization is made using Genetic Algorithm. Based on the results, the optimal inputs are Re = 19,471, the twist pitch of 0.0376, and the volume fraction of 0.0383 producing Nu of 263.57 and a friction factor of 0.0725.

Preliminary results on estimation of signal fading on telecommunication satellite telemetry signals with hybrid numerical weather prediction and artificial neural network approach under presence of aerosol effect

SummaryIn this research, an implementation of artificial deep neural networks (ANN) over outputs of 24‐h multi‐domain high‐resolution nested real case Weather Research and Forecasting (WRF) model runs was carried out over two high‐resolution simulation domains, which are tested and compared for rainfall generation in order to assess the signal fading event observed on geostationary telecommunication spacecraft in orbit for a real multiscale storm case. Our methodology of ANN, which is driven by WRF model output parameters, focuses on prediction of the rain attenuation signal impairment which is observed on the communication satellite telemetry (TM) downlink signal levels under significant aerosol presence due to dust storm which occurred on 12 September 2020. This modelling approach is then compared to rain attenuation observed on TM signal and correlated with communication satellite ground station TM signal measurements. Preliminary results from conducted error analysis (RMSE) on multiple input single output feed‐forward neural network (MISO FFNN) prediction model outputs tested with several neural algorithms indicate good correlation with the TM downlink signal attenuation observations taken from the ground station TM baseband demodulator.