一类完备的正交分段多项式函数系及其应用

在几何图形或图像边界的频谱分析应用中,用Fourier三角基表示间断图形时必然会出现Gibbs现象,而用Walsh函数表示时,因其收敛速度慢而效果欠佳.本文首先构造了一类分段点在四进制有理数点处的分段多项式函数集(简称四进制U-系统,QU-系统),它是L2[0,1]空间上的完备的正交函数系,并研究了它的性质、基函数与Fourier-QU系数的计算公式,同时,也给出了1～3次QU-系统的一组显式表达式.然后,使用Fourier-QU级数的有限项和表示图像轮廓线,提出用有限的Fourier-QU系数描述几何图形或图像轮廓线,并由此得到了一类新的多项式描述子——QU描述子,而归一化QU描述子是一类基于平移、旋转与尺度变换的特征不变量.最后,通过数值实验证实了使用Fourier-QU级数逼近一元平方可积函数时,其收敛速率要优于Fourier级数、Walsh级数和Fourier-BU级数,同样也验证了QU描述子是一类有效的形状描述子,用图像间的QU距离能准确地描述图像间的相似性.     

轮廓特征与神经网络相结合的行人检测

传统的基于方向梯度直方图与支持向量机的行人检测方法运算量大,针对这一问题,本文从轮廓特征的角度出发,提出了头肩轮廓特征与神经网络相结合的检测方法。该方法根据人体头肩模型具有相对稳定性,且轮廓特征可以作为人体识别的依据,采用边缘检测与均值漂移相结合的方式提取人体轮廓,采用经PCA降维的傅里叶描述子提取轮廓特征,结合神经网络分类器完成初次人体识别。采用RGB头发模型和均值漂移方法,对遮挡情况下被判别为非人体的目标图像做进一步处理,聚类出多个人体头肩模型,重新参与分类。实验结果表明,本方法人体检测的准确率和检测速度与现有的算法相比都有所提高,且克服了遮挡情况下人体头肩模型提取错误的弊端,提高了人体检测的识别率和应用范围。     

基于改进模型预测控制的电气系统新能源功率波动平滑策略

针对新能源功率波动问题,制定了聚合温控负荷与微型燃气轮机共同抑制电气互联系统联络线功率波动的调控策略.首先,将波动功率进行经验模态分解,然后在建立聚合温控负荷双线性模型与微型燃气轮机数学模型的基础上,提出了基于Lyapunov函数的改进模型预测控制方法,采用改进模型预测控制分别设计了聚合温控负荷与微型燃气轮机的控制律,...     

A Fragrance Prediction Model for Molecules Using Rough Set-based Machine Learning

In this work, a novel machine learning based methodology was developed to predict fragrance from the molecular structure and the effect of the subjects attributes on odour perception. As fragrance is linked to the molecular structure and interactions, topological indices are used to develop a predictive model. Rough set-based machine learning is used to generate rule-based models that link the topology of fragrant molecules and dilution to their respective odour characteristics. The results show that the generated models are effective in determining the odour characteristic of molecules.     

基于反推控制的聚合温控负荷与发电机励磁协调控制

利用具有储能特性的温控负荷与发电机励磁进行协调控制，可实现电力系统的功角、频率和电压的多指标调控。首先，采用双线性聚合温控负荷模型和发电机励磁模型，推导建立了二者联合的系统数学模型。然后，基于反推控制原理，提出了聚合温控负荷与发电机励磁的协调控制策略，并从理论上证明了控制器的稳定性。针对控制器设计中的微分“计算爆炸”问题，利用双曲正切的非线性微分跟踪器设计了励磁电压参考量的微分估计算法，降低了参考量微分计算的工作量。最后，通过算例对控制算法进行了仿真验证。结果表明，在协调控制器的作用下，系统功角、频率和电压等多个被控量均可快速地追踪上各自的参考值。对比传统电力系统稳定器励磁控制，温控负荷参与调控能有效改善系统各状态量的动态响应，提高系统稳定性。     

基于全景可观的深度聚合电网模型、求解方法及其应用

电力系统调度与控制正面临全局统筹、紧密协同的局面，而对外部电网的等值简化是该决策顺利实施的前提和关键。由于传统等值简化方法对等值电网“源”“网”“荷”的增加或删减操作改变了其结构和主动量、被动量，从而在电网全局统筹过程中，使调度与控制的精度面临挑战。对此，在电网全景可观环境下，提出了深度聚合电网模型及其求解方法和应用。首先，给出了发电厂、联络站的聚合方法，形成聚合电网模型。并在此基础上，将聚合厂站融合到输电线路，建立了深度聚合电网模型。然后，推导了深度聚合电网模型未知量与可调控量(发电机功率)之间的函数关系，并使用长短期记忆网络对其参数进行求解。最后，以经济调度应用为例进行仿真验证，证明了所提方法的有效性。     

FedDAA: a robust federated learning framework to protect privacy and defend against adversarial attack

Federated learning (FL) has emerged to break data-silo and protect clients’ privacy in the field of artificial intelligence. However, deep leakage from gradient (DLG) attack can fully reconstruct clients’ data from the submitted gradient, which threatens the fundamental privacy of FL. Although cryptology and differential privacy prevent privacy leakage from gradient, they bring negative effect on communication overhead or model performance. Moreover, the original distribution of local gradient has been changed in these schemes, which makes it difficult to defend against adversarial attack. In this paper, we propose a novel federated learning framework with model decomposition, aggregation and assembling (FedDAA), along with a training algorithm, to train federated model, where local gradient is decomposed into multiple blocks and sent to different proxy servers to complete aggregation. To bring better privacy protection performance to FedDAA, an indicator is designed based on image structural similarity to measure privacy leakage under DLG attack and an optimization method is given to protect privacy with the least proxy servers. In addition, we give defense schemes against adversarial attack in FedDAA and design an algorithm to verify the correctness of aggregated results. Experimental results demonstrate that FedDAA can reduce the structural similarity between the reconstructed image and the original image to 0.014 and remain model convergence accuracy as 0.952, thus having the best privacy protection performance and model training effect. More importantly, defense schemes against adversarial attack are compatible with privacy protection in FedDAA and the defense effects are not weaker than those in the traditional FL. Moreover, verification algorithm of aggregation results brings about negligible overhead to FedDAA.     

Detection of cardiac amyloidosis on electrocardiogram images using machine learning and deep learning techniques

Cardiac amyloidosis is an uncommon disease that has been known for a long time. Moreover, modern advancement in noninvasive imaging of heart via ultrasound, magnetic resonance imaging has enhanced the detection of secret cardiac amyloidosis in patients identified with the heart disease. This article focused on detecting the heart disease especially cardiac amyloidosis on electro cardio gram images using recent technology of both machine learning and deep learning approaches. In addition, apart from detecting the disease on images, we are categorizing the heart images as normal and cardiac amyloidosis if any deviations occur. For CA disease identification along with its classification, 300 cardiac images have taken and those images are analyzed using machine learning algorithms namely nearest centroid, gradient boosting and random forest. Several metrics such as precision, recall, f-score, sensitivity, accuracy, and confusion matrix based on binary classification which classifies the images into positive (CA) and negative (non-CA) are estimated. Among these approaches, gradient boosting method achieves 95% accuracy as better outcomes which measure the model performance in detecting cardiac amyloidosis disease as well as ECG images are categorized into either normal or abnormal via classification metrics. Furthermore, we applied deep learning based neural network “DeepNet” model is applied on augmented data along with CNN which attains 93% accuracy in CA disease identification.