ID 原文 译文
4353 实验结果表明，所提算法有较好的动态性，结合时间片策略能够获得动态性和异构性的综合平衡，且算法耗时较低 The experimental results show that the algorithm has good dynamics, and itstime slice strategy can achieve a balance of dynamics and heterogeneity, and the algorithm takes less time.
4354 对并行次成分提取算法在信号特征提取中实现方向收敛的过程进行了研究。 The extraction of parallel minor components algorithm in directional convergence in signal feature extractionwas studied.
4355 采用比较未加权的 PAST 次子空间跟踪算法和加权的 PAST 并行次成分提取算法，分析加权规则对次成分提取算法方向收敛行为的限定方式。 By comparing the unweighted projection approximation subspace tracking (PAST) algorithm with theweighted PAST parallel minor component extraction algorithm, the evolution method of the minor component extractionalgorithm was analyzed.
4356 理论分析表明，加权规则对状态矩阵的各个向量和次成分之间的角度限定存在规律，并给出了加权规则对状态矩阵各个向量方向的作用方式和定理和讨论。 Theorical analysis illustrated that the weighted rule was able to guide the angle evolution be-tween the vectors of the state matrix and minor components.
4357 最后，Matlab 仿真实验验证了理论分析的结果。 Finally, Matlab simulation verifies the validity of the pro-posed theory.
4358 针对不同分布噪声下生成对抗网络生成样本质量差异明显的问题，提出了一种噪声稳健性的卡方生成对抗网络。 Aiming at the obvious difference of image quality generated by generative adversarial network under differentnoises, a chi-square generative adversarial network (CSGAN) was proposed.
4359 所提网络结合了卡方散度量化敏感性和稀疏不变性的优势，引入卡方散度计算生成样本分布和真实样本分布的距离，减小不同噪声对生成样本的影响且降低对真实样本的质量要求； Combing the advantages of quantification sensitivity and sparse invariance, the chi-square divergence was introduced to calculate the distance between the generat-ed samples and the original samples, which could reduce the influence of different noises on the generated samples and the quality requirement of original samples.
4360 搭建了网络架构，构建全局优化目标函数，促进网络不断优化并增强博弈的有效性。 Meanwhile, the network architecture was built and the global optimization objective function was constructed to enhance the adversarial performance.
4361 实验结果表明，所提网络在不同噪声下的生成样本质量和稳健性优于目前几种主流网络，且图像质量差异较小。 Experimental results show that the quality of the images generated by the proposed algorithm has little difference, and the network is more robust to different noisesthan the state-of-the-art networks.
4362 卡方散度的引入不仅提高了生成样本质量，而且提升了网络在不同噪声下的稳健性。 The application of chi-square divergence not only improves the quality of generatedimages, but also increases the robustness of the network under different noises.