Stacking-Based Deep Neural Network: Deep Analytic Network for Pattern Classification

Theoretical and experimental research on two-phase flow image reconstruction and flow pattern recognition

Two-phase flow is a kind of complex fluid flow state, and the flow pattern characteristics are very difficult to obtain accurately. First, the principle of two-phase flow pattern image reconstruction based on electrical resistance tomography technology and the complex flow pattern recognition method are developed. Next, the back propagation (BP), wavelet, and radial basis function (RBF) neural networks are applied to the two-phase flow pattern image identification process. The results show that the RBF neural network algorithm has higher fidelity and faster convergence speed than the BP and wavelet network algorithms, and the fidelity is more than 80%. Then, deep learning of the pattern recognition algorithm fusing the RBF network and convolution neural network is proposed to improve the precision of the flow pattern identification. Additionally, the recognition accuracy of the fusion recognition algorithm is more than 97%. Finally, a two-phase flow test system is constructed, the test is finished, and the correctness of the theoretical simulation model is verified. The research process and results provide important theoretical guidance for the accurate acquisition of two-phase flow patterns.

Hybrid modified weighted water cycle algorithm and Deep Analytic Network for forecasting and trend detection of forex market indices

This paper presents forecasting and trend analysis of foreign currency exchange rate in financial market using a hybrid Deep Analytic Network (DAN) technique optimized by a modified water cycle algorithm called Weighted WCA (WWCA) with better generalization capability than the traditional WCA.DAN comprises several stacked KRR (Kernel Ridge Regression) Auto encoders in a multilayer nonlinear regression architecture approach that provides better generalization and accuracy using regularized least squares technique. Further DAN using wavelet kernel function is particularly attractive for its strong data fitting and generalization ability along with its simplified execution procedure, high speed, and better performance achievements in comparison to LSSVM (least squares support vector machine). The output from the DAN is fed to a weighted KRR module to reject noise or the outliers in the noisy data and to make DAN a more robust predictor of the Forex markets, To obtain optimal values of wavelet kernel parameters, a modified metaheuristic water cycle algorithm i.e. the proposed WWCA is utilized. Applications of this new approach to predict forex rate along with trend analysis on three stock markets provide successful results and validate its superiority over some well known approaches like ANN, SVM, Naïve-Bayes, ELM.

Information Fusion Fault Diagnosis Method for Deep-Sea Human Occupied Vehicle Thruster Based on Deep Belief Network

In this article, a novel thruster information fusion fault diagnosis method for the deep-sea human occupied vehicle (HOV) is proposed. A deep belief network (DBN) is introduced into the multisensor information fusion model to identify uncertain and unknown, continuously changing fault patterns of the deep-sea HOV thruster. Inputs for the DBN information fusion fault diagnosis model are the control voltage, feedback current, and rotational speed of the deep-sea HOV thruster; and the output is the corresponding fault degree parameter ( s ), which indicates the pattern and degree of the thruster fault. In order to illustrate the effectiveness of the proposed fault diagnosis method, a pool experiment under different simulated fault cases is conducted in this study. The experimental results have proved that the DBN information fusion fault diagnosis method can not only diagnose the continuously changing, uncertain, and unknown thruster fault but also has higher identification accuracy than the information fusion fault diagnosis methods based on traditional artificial neural networks.

An architecture-level analysis on deep learning models for low-impact computations

Deep neural networks (DNNs) have made significant achievements in a wide variety of domains. For the deep learning tasks, multiple excellent hardware platforms provide efficient solutions, including graphics processing units (GPUs), central processing units (CPUs), field programmable gate arrays (FPGAs), and application-specific integrated circuit (ASIC). Nonetheless, CPUs outperform other solutions including GPUs in many cases for the inference workload of DNNs with the support of various techniques, such as the high-performance libraries being the basic building blocks for DNNs. Thus, CPUs have been a preferred choice for DNN inference applications, particularly in the low-latency demand scenarios. However, the DNN inference efficiency remains a critical issue, especially when low latency is required under conditions with limited hardware resources, such as embedded systems. At the same time, the hardware features have not been fully exploited for DNNs and there is much room for improvement. To this end, this paper conducts a series of experiments to make a thorough study for the inference workload of prominent state-of-the-art DNN architectures on a single-instruction-multiple-data (SIMD) CPU platform, as well as with widely applicable scopes for multiple hardware platforms. The study goes into depth in DNNs: the CPU kernel-instruction level performance characteristics of DNNs including branches, branch prediction misses, cache misses, etc, and the underlying convolutional computing mechanism at the SIMD level; The thorough layer-wise time consumption details with potential time-cost bottlenecks; And the exhaustive dynamic activation sparsity with exact details on the redundancy of DNNs. The research provides researchers with comprehensive and insightful details, as well as crucial target areas for optimising and improving the efficiency of DNNs at both the hardware and software levels.

Prediction of Methylene Blue Removal by Nano TiO2 Using Deep Neural Network

This paper deals with the prediction of methylene blue (MB) dye removal under the influence of titanium dioxide nanoparticles (TiO2 NPs) through deep neural network (DNN). In the first step, TiO2 NPs were prepared and their morphological properties were analysed by scanning electron microscopy. Later, the influence of as synthesized TiO2 NPs was tested against MB dye removal and in the final step, DNN was used for the prediction. DNN is an efficient machine learning tools and widely used model for the prediction of highly complex problems. However, it has never been used for the prediction of MB dye removal. Therefore, this paper investigates the prediction accuracy of MB dye removal under the influence of TiO2 NPs using DNN. Furthermore, the proposed DNN model was used to map out the complex input-output conditions for the prediction of optimal results. The amount of chemicals, i.e., amount of TiO2 NPs, amount of ehylene glycol and reaction time were chosen as input variables and MB dye removal percentage was evaluated as a response. DNN model provides significantly high performance accuracy for the prediction of MB dye removal and can be used as a powerful tool for the prediction of other functional properties of nanocomposites.

English Feature Recognition Based on GA-BP Neural Network Algorithm and Data Mining

With the development of society and the promotion of science and technology, English, as the largest universal language in the world, is used by more and more people. In the life around us, there is information in English all the time. However, because the process of manual recognition of English letters is very labor-intensive and inefficient, the demand for computer recognition of English letters is increasing. This paper studies the influence of the parameters of BP neural network and genetic algorithm on the whole network, including the input, output, and number of hidden layer nodes. Finally, it improves and determines the settings and values of the relevant parameters. On this basis, it shows the rationality of the selected parameters through experiments. The results show that only GA-BP neural network and feature data mining algorithm can complete feature extraction and become the main function of feature classification at the same time. After enough initial data sample analysis training, the GA-BP neural network was found to have good data fault tolerance and feature recognition. The experimental results show that the genetic algorithm can find the best weights and thresholds and the weights and thresholds are given to the BP neural network. After training, the recognition of handwritten letters can be realized. Finally, the convergence of the two algorithms is compared through experiments, which shows that the overall performance of the BP neural network algorithm is improved after genetic algorithm optimization. It can be seen that the genetic algorithm has a good effect in improving the BP neural network and this method has a broad prospect in English feature recognition.

Ensemble Learning with Multiclassifiers on Pediatric Hand Radiograph Segmentation for Bone Age Assessment

In the study of pediatric automatic bone age assessment (BAA) in clinical practice, the extraction of the object area in hand radiographs is an important part, which directly affects the prediction accuracy of the BAA. But no perfect segmentation solution has been found yet. This work is to develop an automatic hand radiograph segmentation method with high precision and efficiency. We considered the hand segmentation task as a classification problem. The optimal segmentation threshold for each image was regarded as the prediction target. We utilized the normalized histogram, mean value, and variance of each image as input features to train the classification model, based on ensemble learning with multiple classifiers. 600 left-hand radiographs with the bone age ranging from 1 to 18 years old were included in the dataset. Compared with traditional segmentation methods and the state-of-the-art U-Net network, the proposed method performed better with a higher precision and less computational load, achieving an average PSNR of 52.43 dB, SSIM of 0.97, DSC of 0.97, and JSI of 0.91, which is more suitable in clinical application. Furthermore, the experimental results also verified that hand radiograph segmentation could bring an average improvement for BAA performance of at least 13%.

Radar Application: Stacking Multiple Classifiers for Human Walking Detection Using Micro-Doppler Signals

We propose a stacking method for ensemble learning to distinguish micro-Doppler signals generated by human walking from background noises using radar sensors. We collected micro-Doppler signals caused by four types of background noise (line of sight (LoS), fan, snow and rain) and additionally considered micro-Doppler signals caused by human walking combined with these four types of background noise. We firstly verified the effectiveness of a fully connected deep neural network (DNN) to classify 8 types of signals. The average accuracy was 88.79% for the test set. Then, we propose a stacking method to combine two base classifiers of different structures. The average accuracy of the stacking method on the test set was 91.43%. Lastly, we designed a modified stacking method to reuse feature information stored at the previous stage and the average test accuracy increased to 95.62%. This result shows that the proposed stacking methods can be an effective approach to improve classifier’s accuracy in recognizing human walking using micro-Doppler signals with background noise.