Feedforward Neural Network Structure Research Articles

Intelligent backpropagated predictive networks for dynamics of the power-law fluidic model with moving wedge and flat plate

A novel application of intelligent computing is presented to study the dynamics of the power-law fluidic problem (PLFP) of moving-wedge and flat-plate models using a multilayer structure of feedforward neural networks (FFNNs) optimized with the Levenberg-Marquardt method (LMM), i.e. FFNN-LMM. The reference dataset for inputs and targets of FFNNs is generated through the Adams numerical solver for the non-Newtonian power-law fluidic system in the case of dilatant, pseudo-plastic, and other thinning/thickening fluids with different values of wedge angle and velocity ratio parameters. The designed FFNNs backpropagated with LMM are constructed with the help of arbitrary samples for training, testing, and validation via mean squared error (MSE)-based figure of merit. Comparison of the proposed computing platform FFNN-LMM with reference standard solutions in terms of MSE, state transition statistics of the algorithm, error histogram studies, and regression outcomes endorse the accuracy and convergence for solving different scenarios in PLFPs.

Using Fuzzy Feed-Forward Neural Network for Linguistic Processing

Fuzzy sets have been implemented efficiently to manage unclear data, language terms, andvague notions. Recently, considerable work has been dedicated to merging neural-networktechniques with fuzzy sets. In this study, present the structure of a fuzzy feed-forward neuralnetwork (FFFNN) with a trapezoidal fuzzy set. In addition to handling real input vectors, it isalso capable of handling fuzzy input vectors. Generally, the output of a FNN is a fuzzy vector.According to the extension principle of Zadeh, each unit of a FNN has an input-outputrelationship. To determine the costs associated with fuzzy calculations and fuzzy objectives,developed a cost function. At that point, created a learning algorithm from the cost capacity toalign the four variables of each trapezoidal fuzzy weight. In conclusion, demonstrate ourmethodology using numerical models.

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>

Residual strength prediction of corroded pipelines using multilayer perceptron and modified feedforward neural network

Corrosion defects occurring in natural gas pipelines are common and annoying. The residual strength prediction of corroded pipelines is usually carried out based on theoretical, numerical, and experimental methods. However, the results are hard to obtain when it comes to high nonlinear problems. In this paper, an artificial neural network (ANN) was used to predicting residual strength of corroded pipelines. Due to inadequate training data from previous experiments and extreme iterations which were needed to ensure precision of predicting results, the overfitting phenomenon occurred. To solve the overfitting phenomenon, the training accuracy was reduced artificially by using ReLU activation function and dropout method which was cutting down neurons during ANN training. The results showed that the multilayer perceptron (MLP) conducted dropout method had the highest precision for inadequate sample data compared with relatively simple feedforward neural network (FFNN) structure and FFNN optimized by particle swarm optimization (PSO).

A new Parkinson detection system based on evolutionary fast learning networks and voice measurements

Parkinson's disease (PD) is becoming the second most neurological syndrome of the central nervous system after Alzheimer's disease. It causes diverse symptoms which include bradykinesia (slowness of movement), voice impairments, rigidity, tremor, and poor balance. PD recognition system based on voice has founded a non-invasive alternative, but involves rather complex measurements or variables. Therefore an attention is required toward new approaches for better forecasting accuracy. In this paper, an optimal fast learning network (FLN) based on genetic algorithm (GA) was established as PD diagnosis system. FLN is a double-parallel feed-forward neural network structure, and based on GA for feature reduction and hyperparameter optimisation of the FLN, it was used as a predictive model. Finally, the conducted experiments on the Parkinson data of voice recordings over ten fold cross-validation show that proposed system is less complex and also achieved better average classification results with an accuracy of 97.47%. At the same time, it is effective in automatic identification of important vocal features. Moreover, the highest average degree of improved accuracy was (2.1%) compared with other familiar wrappers including support vector machine and K-nearest neighbours in the similar conditions.

Handover Optimization Algorithm Based on T2RFS-FNN.

As a key technology for highly reliable communication in the fifth generation mobile communication for railway (5G-R) high-speed railway wireless communication system, once the handover fails, it will pose a serious risk to the safe operation of high-speed railway. As the speed of high-speed trains continues to increase, the handover will become more frequent, and how to improve the success rate of the handover is a key problem that needs to be solved. In this paper, we proposed an optimization algorithm based on the interval type 2 feature selection recurrent fuzzy neural network (T2RFS-FNN), which is a recurrent fuzzy neural network with interval type 2 feature selection, to address the problem of fixed hysteresis threshold and single consideration for the handover algorithm between the control plane and the user plane of the high-speed railway under 5G-R. The algorithm integrates reference signal receiving power (RSRP). Reference signal receiving quality (RSRQ) and throughput to optimise the hysteresis threshold. First, a feedforward neural network structure is designed to implement fuzzy logic inference, and an interval type-two Gaussian subordination function is used to improve the nonlinear expressiveness of the model. Then, a feature selection layer is added to determine the output of the affiliation function, which completes the optimization of the hysteresis threshold and overcomes the drawback of the fixed hysteresis threshold of the handover algorithm. Finally, simulation analysis of the control-plane and user-plane handover algorithms is carried out separately. The results show that the proposed method can effectively improve the success rate and reduce the ping-pong handover rate compared to the comparison algorithms. The results provide a theoretical reference for the speedup of high-speed railway trains and the evolution of the global system for mobile communications for railway (GSM-R) to 5G-R.

DPB-NBFnet: Using neural Bellman-Ford networks to predict DNA-protein binding.

DNA is a hereditary material that plays an essential role in micro-organisms and almost all other organisms. Meanwhile, proteins are a vital composition and principal undertaker of microbe movement. Therefore, studying the bindings between DNA and proteins is of high significance from the micro-biological point of view. In addition, the binding affinity prediction is beneficial for the study of drug design. However, existing experimental methods to identifying DNA-protein bindings are extremely expensive and time consuming. To solve this problem, many deep learning methods (including graph neural networks) have been developed to predict DNA-protein interactions. Our work possesses the same motivation and we put the latest Neural Bellman-Ford neural networks (NBFnets) into use to build pair representations of DNA and protein to predict the existence of DNA-protein binding (DPB). NBFnet is a graph neural network model that uses the Bellman-Ford algorithms to get pair representations and has been proven to have a state-of-the-art performance when used to solve the link prediction problem. After building the pair representations, we designed a feed-forward neural network structure and got a 2-D vector output as a predicted value of positive or negative samples. We conducted our experiments on 100 datasets from ENCODE datasets. Our experiments indicate that the performance of DPB-NBFnet is competitive when compared with the baseline models. We have also executed parameter tuning with different architectures to explore the structure of our framework.

A pruning method of feedforward neural network based on contribution of input components for digital predistortion of power amplifier

AbstractIn this paper, a pruning method based on contribution of input components is proposed to optimize the structure of Feedforward neural network (FNN) model for digital predistortion of power amplifier. The optimized hyperparameters are the specific input components. The pruning algorithm is divided into two stages. The stage I is used to rough rank the importance of input components and the stage II is used to modify the results of stage I. Then, we can get an importance ranking of input components. By deleting unimportant input components, a pruned FNN model with fewer input components can be obtained. Test results show that the proposed algorithm can greatly simplify the FNN model on the premise of ensuring the predistortion performance.

STMMI: A Self-Tuning Multi-Modal Fusion Algorithm Applied in Assist Robot Interaction

While facing complex surroundings, robots need to identify the same intention which is expressed in different ways. In order to solve the problem of assisting robots to get a better intention understanding, a self-tuning multimodal fusion algorithm is put forward in this paper, which is not restricted by the expressions of interacting participants and environment. The multimodal fusion algorithm can be transferred to different application platforms. Robots can own the understanding competence and adapt new tasks by changing the content of the robot knowledge base. Compared with other multimodal fusion algorithms, this paper attempts to transfer the basic structure of feed-forward neural networks on discrete sets, which has strengthened the consistency and perfect the complementary relations between multiple mode, and has realized the simultaneous operation of fusion operator’s self-tuning and intention search. There are three kinds of modes selected in the paper: speech, gesture, and scene objects, where the single modal classifiers are trained separately. This method conducted a human-computer interaction experiment on the bionic robot Pepper platform, which proved that the method can effectively improve the accuracy and robustness of robots in aspects of understanding human intentions, and reduce the uncertainty about intention judgment in a single modal interaction.

A novel sowing operation parameter learning optimization method using dataset of sown seeds with similar properties

A learning optimization method used for the pneumatic roller-type seeder was put forward to determine optimal sowing operation parameters of a new type of seed without conducting performance tests. A feedforward neural network structure with two hidden layers were firstly optimized with Genetic Algorithm (GA) to establish predictive relationships among sowing operation parameters and performance indices. Then particle swarm optimization (PSO) was utilized to search the optimal sowing operation parameters for the new type of seed. Sowing data of ten types of seed from previous study with the 2BS-6 pneumatic roller seeder were gathered as dataset of sown seeds. Cosine similarity of seed physical properties was analyzed to extract training datasets for eggplant and mustard. The lowest R-value of 0.85 and the largest MAE value of 4.38 were obtained by regression analysis of the eggplant and mustard predictive model. The smallest R2 of 0.731of optimized sowing performance indices were reported. Optimized fitness values were 83.15% for eggplant and 87.44% for mustard. Fitness value deviations of eggplant and mustard were 6.51% and 1.58% respectively. For eggplant, experimental deviations of single-seeding rate, multi-seeding rate and miss-seeding rate were 2.1%, 2.39% and 0.29% respectively. All experimental deviations of mustard were less than that of eggplant. The study demonstrated high predictive accuracy, strong optimal ability, and promising applicability of the proposed method to determine sowing operation parameters efficiently for the pneumatic roller-type seeder.

Improved Artificial Neural Network Training Based on Response Surface Methodology for Membrane Flux Prediction.

This paper presents an improved artificial neural network (ANN) training using response surface methodology (RSM) optimization for membrane flux prediction. The improved ANN utilizes the design of experiment (DoE) technique to determine the neural network parameters. The technique has the advantage of training performance, with a reduced training time and number of repetitions in achieving good model prediction for the permeate flux of palm oil mill effluent. The conventional training process is performed by the trial-and-error method, which is time consuming. In this work, Levenberg–Marquardt (lm) and gradient descent with momentum (gdm) training functions are used, the feed-forward neural network (FFNN) structure is applied to predict the permeate flux, and airflow and transmembrane pressure are the input variables. The network parameters include the number of neurons, the learning rate, the momentum, the epoch, and the training functions. To realize the effectiveness of the DoE strategy, central composite design is incorporated into neural network methodology to achieve both good model accuracy and improved training performance. The simulation results show an improvement of more than 50% of training performance, with less repetition of the training process for the RSM-based FFNN (FFNN-RSM) compared with the conventional-based FFNN (FFNN-lm and FFNN-gdm). In addition, a good accuracy of the models is achieved, with a smaller generalization error.

Design of FFNN architecture for power quality analysis and its complexity challenges on FPGA

As we all know, power quality (PQ) issues are a major concern these days. Field programmable gate array (FPGA) are essential in PQ analysis, particularly in smart meters for data processing, storage, and transmission. One of the most significant advantages of FPGA is its reconfigurability, with vast hardware resources that can be used to implement complex as well as time-critical data processing units. Because the FPGA architecture supports fixed point arithmetic, data loss occurs in the data path unit, necessitating the realization of the PQ event detection module, and classification model to be more accurate than software implementation algorithms. The majority of the work reported, with feed forward neural network (FFNN) structure occupying large number of multipliers and adders for classification, most of the work reported has not addressed to minimize the data path resources for FFNN instead have addressed in improving classification accuracy. Based on these issues, this paper addresses the implementation challenges in FFNN architecture design by proposing improved and fast architectures. The proposed FFNN architecture design using optimum resources. The FFNN based classifier are designed to perform PQ event detection and classification with 99.5% accuracy. The FFNN processor operates at maximum frequency of 238 MHz.

Empowering saving energy at home through serious games on thermostat interfaces

The residential Heating Ventilation and Air-Conditioning (HVAC) system use around 3/5 of the total energy consumption. Connected thermostats optimize the HVAC operation; however, householders have personality traits that lead into behavioral and usability problems toward the thermostat’s interface usage. Thus, a serious game applied in the thermostat interface can balance entertainment and education. Therefore, thermostat interfaces must address strategies that reduce energy without losing thermal comfort. This paper proposed an interactive interface type and a predicted interface type based on an HVAC strategy and a Natural Ventilation strategy. These strategies measured the impact of adaptive thermal comfort, energy consumption, and costs. Hence, twelve energy models located in California (Concord, Riverside, Los Angeles, and San Diego) were simulated using EnergyPlus™ through LadybugTools. The first interactive interface included Serious Game elements, so the householder interacted with the date, location, and setpoint. The second interface predicted the energy consumption and thermal comfort during winter and summer in Concord by a two-layer feed-forward Artificial Neural Network structure. The proposed structure decreases the energy consumption by at least 62% without losing thermal comfort.

Global Pattern Feedforward Neural Network Structure with Bacterial Foraging Optimization towards Medicinal Plant Leaf Identification and Classification

Medicinal Plant species help to cure various diseases across the world. The automated identification of medicinal plant species to treat disease based on their structure is much required in pharmaceutical laboratories. Plant Species with a complex background in the field will make the detection and classification more difficult. In this paper, optimization of bacterial foraging technique has been employed towards medicinal plant prediction and classification architecture based on feed-forward neural network. It is capable of identifying both complex structures of medicinal plants. Feed-forward Neural Networks are considered to have good recognition accuracy compared to other machine learning approaches. Further bacterial foraging has been implemented to minimize the feature search space to the classifier and provides optimal features for the plant classification. The experimental outcomes of the proposed approach has been analysed by employing the medley dataset and evaluating the performance of the proposed approach with respect to dice similarity coefficient, Specificity and sensitivity towards medicinal plant classification. The findings are very positive, and further research will focus on using a large dataset and increased computing resources to examine how well deep-learning neural networks function in identifying medicinal plants for use in health care.

Modelling the hardened properties of steel fiber reinforced concrete using ANN

The hardened properties of concrete are one of the most important parameters when considering the sustainability criteria of buildings. Numerous previous studies have established how the Artificial Neural Network is used as an effective statistical data modelling tool. In this study, the prediction of Elastic modulus and Compressive Strength of steel fibre reinforced concrete have been illustrated with a feed-forward backpropagation neural network structure. 158 datasets are used for modelling of elastic modulus and 140 datasets is used for modelling of Compressive Strength. The developed ANN models were able to predict the data within the range of input parameters considered and were having regression coefficients values of 0.96 and 0.97 respectively. A comparison of actual and predicted values has been performed and the results indicated that the ANN performed better in terms of prediction with minimum deviation from original data.

Estimation of power output and thermodynamic analysis of standard and finned photovoltaic panels

ABSTRACT This study deals with the estimation of power output and also thermodynamic analysis of two different photovoltaic panels. One of the panels is a standard photovoltaic module without fins (SPV), and the other one is a photovoltaic module with fins (FPV). First, a multi-layer feed-forward neural network structure is designed to estimate the daily power produced by photovoltaic modules. Furthermore, energy and exergy analyses were carried out to compare the performance of SPV and FPV panels. According to the thermodynamic analysis results using the experimental data obtained for two days (July 3, 2020 and August 4, 2020), it was calculated that the energy efficiency increased by a maximum of 8.77% and the exergy efficiency increased by a maximum of 25.9% in the FPV panel compared to the SPV panel. Moreover, considering the data obtained for each day during three months (July, August, and September), the total energy production increase in the FPV panel is approximately 6.7% compared to the SPV panel.

Optimizing Deep Learning Methods in Neural Network Architectures

Deep neural networks are a powerful tool for computer-assisted learning and have achieved significant success in numerous computer vision and image processing tasks. This paper discusses several new neural network structures that have better performance than the traditional feedforward neural network structure. A method of network structure optimization based on gradient descent and heavy-ball algorithms has been proposed. Furthermore, an approach based on the concept of sparse representation for simultaneous training and optimizing the network structure has been presented. According to CIFAR-10 and CIFAR-100 dataset classification tasks and experimental results, the optimization of ResNet and DenseNet structures using gradient descent and heavy-ball algorithms, accordingly, has been shown to result in better performance with increased depth of neural network. A neural network based on a sparse representation is shown to have the highest performance in all datasets. This strategy encourages quick data adaptation at each iteration. The results obtained can be used to design deeper neural networks with no loss of precision and computing speed.

Shale gas well flowback rate prediction for Weiyuan field based on a deep learning algorithm

The flowback rate of a shale gas well is controlled by many factors, such as geological and engineering factors, which have a certain complexity. For the aforementioned reason, the knowledge-driven method often has difficulty effectively capturing the correlation rules among multiple datasets. As a data-driven algorithm, deep learning has strong advantages in data correlation analysis, nonlinear fitting and other applications. In this paper, an algorithm is set up to forecast the shale gas well flowback rate by using a deep learning algorithm based on the flowback characteristic factors and data from 286 shale gas wells in the Weiyuan field. An algorithm and prediction model of the flowback parameters of the Weiyuan shale gas wells is established to effectively capture their flowback characteristics. First, the general situation of the study area is briefly introduced, and the data used to predict the flowback rate in this area are analyzed. Second, based on the deep learning algorithm, two kinds of deep feedforward neural network structures and a neural network-based prediction method of shale gas well flowback rate are designed. Finally, according to the regional data, the prediction and analysis of the flowback rate of Weiyuan shale gas wells are carried out. The results show that the correlations between 17 factors, such as fracturing section length and shale gas well flowback rate, are not obvious, but that the deep learning method proposed in this paper can effectively and accurately capture the correlations between the data and predict the flowback rate. The results show high accuracy.

An All-Batch Loss for Constructing Prediction Intervals

The prediction interval (PI) is an important research topic in reliability analyses and decision support systems. Data size and computation costs are two of the issues which may hamper the construction of PIs. This paper proposes an all-batch (AB) loss function for constructing high quality PIs. Taking the full advantage of the likelihood principle, the proposed loss makes it possible to train PI generation models using the gradient descent (GD) method for both small and large batches of samples. With the structure of dual feedforward neural networks (FNNs), a high-quality PI generation framework is introduced, which can be adapted to a variety of problems including regression analysis. Numerical experiments were conducted on the benchmark datasets; the results show that higher-quality PIs were achieved using the proposed scheme. Its reliability and stability were also verified in comparison with various state-of-the-art PI construction methods.

Vortex search optimization algorithm for training of feed-forward neural network

Training of feed-forward neural-networks (FNN) is a challenging nonlinear task in supervised learning systems. Further, derivative learning-based methods are frequently inadequate for the training phase and cause a high computational complexity due to the numerous weight values that need to be tuned. In this study, training of neural-networks is considered as an optimization process and the best values of weights and biases in the structure of FNN are determined by Vortex Search (VS) algorithm. The VS algorithm is a novel metaheuristic optimization method recently developed, inspired by the vortex shape of stirred liquids. VS fulfills the training task to set the optimal weights and biases stated in a matrix. In this context, the proposed VS-based learning method for FNNs (VS-FNN) is conducted to analyze the effectiveness of the VS algorithm in FNN training for the first time in the literature. The proposed method is applied to six datasets whose names are 3-bit XOR, Iris Classification, Wine-Recognition, Wisconsin-Breast-Cancer, Pima-Indians-Diabetes, and Thyroid-Disease. The performance of the proposed algorithm is analyzed by comparing with other training methods based on Artificial Bee Colony Optimization (ABC), Particle Swarm Optimization (PSO), Simulated Annealing (SA), Genetic Algorithm (GA) and Stochastic Gradient Descent (SGD) algorithms. The experimental results show that VS-FNN is generally leading and competitive. It is also said that VS-FNN can be used as a capable tool for neural networks.