Hyperparameter on-line learning of stochastic resonance based threshold networks

Aiming at training the feed-forward threshold neural network consisting of nondifferentiable activation functions, the approach of noise injection forms a stochastic resonance based threshold network that can be optimized by various gradient-based optimizers. The introduction of injected noise extends the noise level into the parameter space of the designed threshold network, but leads to a highly non-convex optimization landscape of the loss function. Thus, the hyperparameter on-line learning procedure with respective to network weights and noise levels becomes of challenge. It is shown that the Adam optimizer, as an adaptive variant of stochastic gradient descent, manifests its superior learning ability in training the stochastic resonance based threshold network effectively. Experimental results demonstrate the significant improvement of performance of the designed threshold network trained by the Adam optimizer for function approximation and image classification.     

一类中立型Cohen-Grossberg神经网络的概周期解

研究一类具有混合时滞的中立型Cohen-Grossberg神经网络。通过建立线性辅助方程， 得到该神经网络存在唯一的概周期解的新结果，同时也给出此概周期解的存在范围。     

考虑多种损坏构成特征的沥青路面预防性养护决策方法

由于沥青路面损坏构成的多样性, 相同的路面状况指数(pavement condition index, PCI)可能代表不同的损坏组合. 当多种损坏并存且损坏程度接近时, 用PCI和主导损坏(最严重、扣分最多的路面损坏)难以得到具有针对性的养护对策. 因此, 通过对PCI的深入分析, 明确了主导损坏代表性不足的路段, 以现行预防性养护决策方法为基础, 补充了一种考虑损坏构成特征、更具针对性的决策方法. 以上海城市道路近5年的检测、养护数据为分析基础, 首先利用有序聚类算法将路段按PCI水平分组, 分析了不同阶段路面损坏构成和差异水平; 然后, 针对多种损坏并存且损坏差异不显著的路段, 根据预防性养护的实施效果筛选了能够反映正确预防性养护经验的有效养护路段; 最后, 基于有效养护路段建立并对比分析了2个基于BP(back propagation)神经网络的养护决策模型. 结果表明: 当PCI水平介于优良(84.4~93.0分)时, 不同损坏程度接近, 主导损坏代表性不足; 考虑多种损坏构成特征的BP神经网络模型表现出更高的决策精度, 测试集决策正确率达86.20%, 优于仅考虑主导损坏的模型(58.50%). BP神经网络与传统决策树法结合能够优化沥青路面决策过程, 提高养护对策选取的针对性.     

一类离散时间复杂网络系统的输出同步

利用线性稳定性理论和Lyapunov函数方法,研究了一类离散时间复杂网络系统的输出同步问题,得到了系统输出指数同步的充分性条件,数值算例表明了方法的有效性.     

植入式硅神经微电极的发展

 神经科学和神经工程研究需要研究大脑神经元的电活动情况，以了解大脑产生、传输和处理信息的机制。植入式神经微电极作为一种传感器件，是时间分辨率最高的神经电活动传感手段之一。介绍了国内外几种主要的植入式硅基神经微电极的结构特点、制备方法和性能特点。分析表明，未来通过不断结构优化和改性修饰，特别是在高通量的神经记录方面，通过与同样基于硅材料的电路的集成，硅神经微电极能够进一步提高生物相容性，解决大规模的电极通道体内外传输与连接问题，实现对神经元的在体大规模长时间记录。     

基于免疫粒子群优化的移动Sink路径规划算法

为了减少能量空洞和延长网络生命周期，在无线传感网中采用移动 Sink 的方式收集节点采集的数据是解决能 量效率问题的有效措施．采集路径的规划问题类似于旅行商问题，无法得到多项式时间的解．提出了将人工免疫算法和粒子群算法相结合，针对移动 sink 数据收集的路径规划问题寻求近似最优解，仿真结果表明: 与其他算法进行性能比较，所提出的优化算法能够有效减少能耗和缩短遍历路径．     

Current air quality analytics and monitoring: A review

This review summarizes the different tools and concepts that are commonly applied in air quality monitoring. The monitoring of atmosphere is extremely important as the air quality is an important problem for large communities. Main requirements for analytical devices used for monitoring include a long period of autonomic operation and portability. These instruments, however, are often characterized by poor analytical performance. Monitoring networks are the most common tools used for monitoring, so large-scale monitoring programmes are summarized here. Biomonitoring, as a cheap and convenient alternative to traditional sample collection, is becoming more and more popular, although its main drawback is the lack of standard procedures. Telemonitoring is another approach to air monitoring, which offers some interesting opportunities, such as ease of coverage of large or remote areas, constituting a complementary approach to traditional strategies; however, it requires huge costs.     

基于变量选择-神经网络模型的复杂路网短时交通流预测

摘要： 针对传统交通流预测模型正在由单断面历史数据处理向多断面、多时刻历史数据处理转变,但在考虑各断面间的影响时,多变的交通状况往往会使预测模型复杂化的问题,引入一种多元线性回归最小绝对收缩和选择算子方法(Lasso),并利用其优秀的变量选择能力,在复杂路网多断面中选出相关性较高的断面；结合神经网络(NN)的非线性特性,提出了Lasso NN组合模型.结果表明：Lasso NN模型在路网交叉口对未来15 min交通流数据预测的误差率低于9.2%；在非交叉口的误差率低于6.7%,总体优于各自单独使用得出的结果.