Polarimetric dynamic vision sensor p(DVS) neural network architecture for motion classification

With the development of the wind power industry, wind turbine blades are increasingly adopting ultra-large-scale designs. However, as the size of blades continues to increase, existing aerodynamic calculation methods struggle to achieve both relatively high computational accuracy and efficiency simultaneously. To tackle this challenge, this research focuses on the low accuracy issues of the traditional Blade Element Momentum theory (BEM) in predicting the aerodynamic performance of wind turbine blades. Consequently, a correction framework is proposed, to integrate the Computational Fluid Dynamics (CFD) method with the Multilayer Perceptron (MLP) neural network. In this approach, the CFD method is used to predict the airflow characteristics around the blades, and the MLP neural network is employed to model the intricate functional relationships between multiple influencing factors and key aerodynamic parameters. This process results in high-precision predictive functions for key aerodynamic parameters, which are then used to correct the traditional BEM. When this correction framework is applied to the rotor of the IEA 15 MW wind turbine, the effectiveness of MLP in predicting key aerodynamic parameters is demonstrated. The research findings suggest that this framework can enhance the accuracy of BEM aerodynamic load predictions to a level comparable to that of RANS.

This article considers the impact of different thresholding methods to the Nearest Shrunken Centroid algorithm, which is popularly referred as the Prediction Analysis of Microarrays (PAM) for high-dimensional classification. PAM uses soft thresholding to achieve high computational efficiency and high classification accuracy but in the price of retaining too many features. When applied to microarray human cancers, PAM selected 2611 features on average from 10 multi-class datasets. Such a large number of features make it difficult to perform follow up study. One reason behind this problem is the soft thresholding, which is known to produce biased parameter estimate in regression analysis. In this article, we extend the PAM algorithm with two other thresholding methods, hard and order thresholding, and a deep search algorithm to achieve better thresholding parameter estimate. The modified algorithms are extensively tested and compared to the original one based on real data and Monte Carlo studies. In general, the modification not only gave better cancer status prediction accuracy, but also resulted in more parsimonious models with significantly smaller number of features.

Recent advances in the machine learning field, especially in deep learning, provide the opportunity for automated, detailed, and unbiased analysis of motor behavior. Although there has not yet been wide use of these techniques in the motor rehabilitation field, they have great potential. In this review, I describe how the current state of machine learning can be applied to 3D kinematic analysis, and how this will have an impact on neurorehabilitation. Applications of deep learning methods, in the form of convolutional neural networks, have been revolutionary for image analysis such as face recognition and object detection in images, exceeding human level performance. Recent studies have shown applicability of these deep learning approaches to human posture and movement classification. It is to be expected that portable stereo-camera systems will bring 3D pose estimation into the clinical setting and allow the assessment of movement quality in response to interventions. Advances in machine learning can help automate the process of obtaining 3D kinematics of human movements and to identify/classify patterns of movement.

To further improve the efficiency of existing sports quality evaluation methods and enhance the granularity of results, this study proposes an intelligent sports quality evaluation system based on Generative Artificial Intelligence (GAI). The system first includes a spatiotemporal motion reconstruction branch based on a Conditional Generative Adversarial Network. Meanwhile, it introduces a Gated Recurrent Unit with an attention mechanism to capture long-term temporal dependencies, thus generating high-quality motion reconstruction sequences. Second, the study uses an error detection and interpretation branch of the Variational Autoencoder (VAE), which separates style and content features through latent space analysis to accurately locate and quantify error patterns. The two branches are jointly optimized through a Dynamic Weighting Fusion (DWF) mechanism, and finally output evaluation results containing multi-dimensional quantitative indicators such as Joint Angle Error (JAE) and motion trajectory error. To fully verify the system’s performance in complex real-world scenarios, this study is based on the Physical Activity Monitoring for Age Group 2 (PAMAP2) dataset to construct a benchmark dataset named PAMAP2-Motion Enhanced (PAMAP2-ME). This dataset introduces simulated real motion noise, multi-view motion sequences, and common incorrect motion patterns demonstrated by professional coaches through GAI technology. Experimental results show that the proposed model’s motion classification accuracy is 96.2%, which is 3.7% higher than that of the baseline model, Multi-Scale Gated 3D Convolutional Network (MS-G3D). In terms of fine-grained evaluation, the model also performs better than MS-G3D in indicators such as JAE (4.5°) and Mean Squared Error (MSE) of motion trajectory reconstruction (0.032). The above data indicate that GAI technology can improve the evaluation system’s efficiency and feedback granularity, providing reliable technical support for intelligent sports teaching and training.

An interpretable neural network (NN) is used to model the self-heating (SH) in complex FinFET circuits. The NN training/testing datasets from 3 -stage to 37 -stage chain circuits in folded layout composed of inverter (INV)/NAND/NOR are simulated by our distributed $\mathrm{R}_{\mathrm{th}}-\mathrm{C}_{\mathrm{th}}$ SPICE model [1], [2]. The interfacial thermal resistance [3], boundary scattering [4], alloy scattering [5], and layout dependence are considered. The NN interpretation by feature importance analysis is consistent with the thermal physics. Stage# is the most important feature of the NN prediction. Both via2 bundle positions and via2 numbers (via2#) are effective to reduce SH. As compared to SPICE, NN prediction in 37 -stage INV chain computes $3\mathrm{x}10^{6}\mathrm{X}$ faster with accuracy loss $ . The high computation efficiency and high precision make NN feasible to predict chain circuits up to 40 stages, which cannot be simulated by SPICE due to long computation time.

Classification and quantitative aspects of fetal movements during the second half of normal pregnancy

Polarimetric fiber optic sensors have many advantages over other fiber optic sensors and strain gauges. As opposed to Fiber Bragg Grating sensors, polarimetric fiber sensors can be constructed with very simple, portable and robust experimental setup. In the paper we propose two methods of the polarimetric sensor output signal analyzing. First method is based on depolarization phenomena, and calculates degree of polarization (DOP), second one uses Fast Fourier Transform to analyze dynamic sensor response. Usefulness of both methods has been shown and experimentally proven. Full Text: PDF References K. Schroeder, W. Ecke, J. Apitz, E. Lembke, G. Lenschow, "A fibre Bragg grating sensor system monitors operational load in a wind turbine rotor blade," Meas. Sci. Technol. 17, 1167–1172 (2006). CrossRef G. Hegde, A. Asundi, "Performance analysis of all-fiber polarimetric strain sensor for composites structural health monitoring," NDT&E International 39, 320–327 (2006). CrossRef G. Wehrle, P. Nohama, H. J. Kalinowski, P. I. Torres, L. C. G.Valente, "A fibre optic Bragg grating strain sensor for monitoring ventilatory movements" Meas. Sci. Technol. 12, 805–809 (2001). CrossRef W. J. Bock, A. W. Domanski, T. R. Wolinski, "Influence of High Hydrostatic Pressure on Beat Length in Highly Birefringent Single-Mode Bow-Tie Fibers," Appl. Optics 29, 3484-3488 (1990). CrossRef A. W. Domański, M. Bieda, P. Lesiak, P. Makowski, M. Szeląg, T. Poczęsny, K. Prokopczuk, P. Sobotka, M. Chychłowski, M. Sierakowski, T. R. Woliński, "Polarimetric optical fiber sensors for dynamic strain measurement in composite materials," Acta Physica Polonica A. 124, 399-401 (2013). CrossRef P. Lesiak, G. Rajan, Y. Semenova, G. Farrell, A. Boczkowska, A. W. Domański, T. R. Woliński, "A hybrid highly birefringent fiber optic sensing system for simultaneous strain and temperature measurement," PLP 3, 140-142 (2010). CrossRef R. A. Chipman, "Depolarization in the Mueller Calculus," Proc. SPIE 5158, 184–192 (2003). CrossRef N. Furstenau, M. Schmidt, W. J. Bock, W. Urbanczyk, "Dynamic pressure sensing with a fiber-optic polarimetric pressure transducer with two-wavelength passive quadrature readout," App. Opt. 37, 663-671 (1998). CrossRef

This work is aimed at developing intelligent systems capable of automatically classifying types of educational materials. This will allow students to find the resources they need faster, and it will make it easier for teachers to manage content in educational platforms. The solution of the problem of recognition of information objects using fuzzy output systems and neural networks is considered. This approach combines the advantages of neural machine learning with the flexibility and efficiency of fuzzy logic, making these systems effective tools for solving problems related to fuzzy or uncertain data. An information model of a neural network for classifying information objects in e-learning systems has been developed. Experimental testing of the proposed approach was carried out on a data set, which consists of information objects from real e-learning systems, namely such as manuals, lectures, syllabuses, and textbooks. Design an adaptive mechanism that utilizes fuzzy neural networks to personalize content recommendations for students based on their learning progress and preferences, thereby enhancing the effectiveness of e-learning systems. Improve the efficiency of classification algorithms by fine-tuning neural network parameters and fuzzy logic rules, ensuring high accuracy and computational efficiency when processing large-scale educational datasets. The results of experimental studies have shown that a neural network built based on fuzzy logic is able to classify various information objects efficiently and correctly in e-learning systems. It is shown that the integration of neural networks based on fuzzy logic into e-learning systems to improve the processes of classification of information objects makes it possible to increase the efficiency of educational resources management, ensuring accuracy and flexibility in processing various data.

Forced vibration analysis of thin cross-ply laminated circular cylindrical shells with arbitrary boundary conditions using the symplectic wave-based method

Free vibration of the complex cable system − An exact method using symbolic computation

We provide a unifying framework where artificial neural networks and their architectures can be formally described as particular cases of a general mathematical construction---machines of finite depth. Unlike neural networks, machines have a precise definition, from which several properties follow naturally. Machines of finite depth are modular (they can be combined), efficiently computable, and differentiable. The backward pass of a machine is again a machine and can be computed without overhead using the same procedure as the forward pass. We prove this statement theoretically and practically via a unified implementation that generalizes several classical architectures---dense, convolutional, and recurrent neural networks with a rich shortcut structure---and their respective backpropagation rules.

Based on the improved complex variable moving least-squares (ICVMLS) approximation, the improved complex variable element-free Galerkin (ICVEFG) method for the bending problem of Kirchhoff plate is presented. Compared with the moving least-squares (MLS) approximation, in the ICVMLS approximation, the approximation function of two-dimensional problems can be obtained with one-dimensional basis function, then the computational efficiency of the shape functions is higher. Compared with the meshless methods based on the MLS approximation, under the same node distributions, the ones using the ICVMLS approximation can obtain the solutions with higher computational accuracy; and under the similar computational accuracy, the ones using the ICVMLS approximation have higher computational efficiency. The ICVMLS approximation is used to form the approximation function of the deflection of a Kirchhoff plate, the Galerkin weak form of the bending problem of Kirchhoff plates is adopted to obtain the discretized system equations, and the penalty method is employed to enforce the essential boundary conditions, then the corresponding formulae of the ICVEFG method for the bending problem of Kirchhoff plates are presented. By computing and analyzing four typical examples, it is shown that the ICVEFG method of Kirchhoff plates in this paper is efficient. And the computational precision of the numerical solutions is analyzed to select the basis function, weight function, scaling factor, node distribution and penalty factor in the ICVEFG method. Numerical examples show that the method in this paper has better convergence and higher accuracy. When the quadratic polynomial basis function and the cubic or quartic spline weight function are used, and d m a x = 4.2−4.4 , the ICVEFG method of Kirchhoff plates in this paper can obtain the solutions with high computational accuracy. And more nodes are distributed in the problem domain, higher computational accuracy of the solution will obtained, which shows that the method in this paper has better convergence.

The complex damping model can only be used to calculate the steady-state responses, while the transient responses are divergent. Based on the complex damping model, the Hilbert transform is introduced to establish a hysteretic damping model, eliminating the divergence phenomenon. However, with the increase of the loss factor, the damped natural frequency also increases. To overcome this shortcoming, a frequency-independent damping model is proposed based on the hysteretic damping model. However, traditional time-domain methods are no longer applicable to frequency-independent damping models. Therefore, the transient response and steady-state response are separated, and the assumption of external excitation acceleration is introduced. Time-domain methods-based linear polynomial assumption, quadratic polynomial assumption and trigonometric function assumption are proposed, respectively. Numerical examples show that the time-domain methods based on linear polynomial assumption and quadratic polynomial assumption have high computational efficiency. But these two methods cannot take into account the vibration frequency of external excitation acceleration. Hence, the computational accuracy is low. Compared with them, the time-domain method based on trigonometric function assumption has the lowest computational efficiency and the highest computational accuracy.

Quantum machine learning (QML) has emerged as a promising approach for enhancing image classification by exploiting quantum computational principles such as superposition and entanglement. However, practical applications on complex datasets like CIFAR-100 remain limited due to the low expressivity of shallow circuits and challenges in circuit optimization. In this study, we propose HQCNN–REGA—a novel hybrid quantum–classical convolutional neural network architecture that integrates data re-uploading and genetic algorithm optimization for improved performance. The data re-uploading mechanism allows classical inputs to be encoded multiple times into quantum states, enhancing the model’s capacity to learn complex visual features. In parallel, a genetic algorithm is employed to evolve the quantum circuit architecture by optimizing gate sequences, entanglement patterns, and layer configurations. This combination enables automatic discovery of efficient parameterized quantum circuits without manual tuning. Experiments on the MNIST and CIFAR-100 datasets demonstrate state-of-the-art performance for quantum models, with HQCNN–REGA outperforming existing quantum neural networks and approaching the accuracy of advanced classical architectures. In particular, we compare our model with classical convolutional baselines such as ResNet-18 to validate its effectiveness in real-world image classification tasks. Our results demonstrate the feasibility of scalable, high-performing quantum–classical systems and offer a viable path toward practical deployment of QML in computer vision applications, especially on noisy intermediate-scale quantum (NISQ) hardware.

Event-driven sorting algorithm-based Monte Carlo method with neighbour merging method for solving aerosol dynamics