基于精确时间协议的HL-2M装置定时同步网络研究

基于HL-2M装置对时序精度的要求,参考ITER的设计方案,设计了基于精确时间协议(PTP)的HL-2M分布式时间通讯网络,使HL-2M装置的时钟同步和事件触发的精度从微秒级提高到亚微秒级,最终优于100ns.     

复杂网络上的意见动力学对谣言传播的影响

通过在SIR(susceptible-infected-recovered)模型中引入抑制者对谣言的辟谣机制研究了在线社交网络上的意见动力学对谣言传播的影响.在这一模型中,节点可以与自身的邻居组成1个群,传播者可以通过该群传播信息,抑制者也可以在此群中对信息发表意见进行辟谣.辟谣机制在降低未知者对于谣言的接受概率的同时也可以促使传播者向抑制者转变.本文采用ER(Erd?s-Rényi)随机网络、无标度网络以及真实的社交网络研究了抑制者的沉默概率对于谣言传播范围的影响.首先发现,谣言传播的过程以传播者的峰值为界可以分为两个阶段,即谣言自由传播的前期以及抑制者和传播者互相制衡的后期;其次,谣言的传播会随着抑制者的沉默概率的增大而突然暴发.在谣言暴发阈值之下,沉默概率的增大不会导致谣言传播范围显著增大,这是由于未知者在感知到谣言并转变为传播者后又迅速转变为抑制者;而当沉默概率达到谣言暴发阈值时,抑制者将不能控制传播者对谣言的传播从而导致抑制者的降低和谣言的暴发;最后,无标度上的谣言自由传播的前期阶段比随机网络持续的时间更短,从而使无标度上的谣言更难以暴发.本文的模型综合考虑了意见动力学和谣言传播的相互作用,更加真实地模拟了真实世界社交网络中的谣言传播过程.为谣言传播的控制和干预提供了一些有用的思路和见解.     

Network-Based Driving Force of National Economic Development: A Social Capital Perspective

Network science has been widely applied in theoretical and empirical studies of global value chain (GVC), and many related articles have emerged, forming many more mature and complete analytical frameworks. Among them, the GVC accounting method based on complex network theory is different from the mainstream economics in both research angle and content. In this paper, we build up global industrial value chain network (GIVCN) models based on World Input–Output Database, introduce the theoretical framework of Social Capital, and define the network-based indicators with economic meanings. Second, we follow the econometric framework to analyze the hypothesis and test whether it is true. Finally, we study how the three types of capital constituted by these indicators interact with each other, and discuss their impact on the social capital (economic development level, i.e., GDP). The results prove that the structural capital (industrial status) has a positive impact on the social capital; the relational capital (industrial correlation) has a positive impact on both social capital and structural capital; the cognitive capital (industrial structure) has a small impact on the social capital, structural capital, and relational capital.     

(Semi-)Automatically Parsing Private Protocols for In-Vehicle ECU Communications

In-vehicle electronic control unit (ECU) communications generally count on private protocols (defined by the manufacturers) under controller area network (CAN) specifications. Parsing the private protocols for a particular vehicle model would be of great significance in testing the vehicle’s resistance to various attacks, as well as in designing efficient intrusion detection and prevention systems (IDPS) for the vehicle. This paper proposes a suite of methods for parsing ECU private protocols on in-vehicle CAN network. These methods include an algorithm for parsing discrete variables (encoded in a discrete manner, e.g., gear state), an algorithm for parsing continuous variables (encoded in a continuous manner, e.g., vehicle speed), and a parsing method based on upper-layer protocols (e.g., OBD and UDS). Extensive verifications have been performed on five different brands of automobiles (including an electric vehicle) to demonstrate the universality and the correctness of these parsing algorithms. Some parsing tips and experiences are also presented. Our continuous-variables parsing algorithm could run in a semi-automatic manner and the parsing algorithm from upper-layer protocols could execute in a completely automatic manner. One might view the results obtained by our parsing algorithms as an important indicator of penetration testing on in-vehicle CAN network.     

Minimum Message Length in Hybrid ARMA and LSTM Model Forecasting

Modeling and analysis of time series are important in applications including economics, engineering, environmental science and social science. Selecting the best time series model with accurate parameters in forecasting is a challenging objective for scientists and academic researchers. Hybrid models combining neural networks and traditional Autoregressive Moving Average (ARMA) models are being used to improve the accuracy of modeling and forecasting time series. Most of the existing time series models are selected by information-theoretic approaches, such as AIC, BIC, and HQ. This paper revisits a model selection technique based on Minimum Message Length (MML) and investigates its use in hybrid time series analysis. MML is a Bayesian information-theoretic approach and has been used in selecting the best ARMA model. We utilize the long short-term memory (LSTM) approach to construct a hybrid ARMA-LSTM model and show that MML performs better than AIC, BIC, and HQ in selecting the model—both in the traditional ARMA models (without LSTM) and with hybrid ARMA-LSTM models. These results held on simulated data and both real-world datasets that we considered.We also develop a simple MML ARIMA model.     

Whether the Support Region of Three-Bit Uniform Quantizer Has a Strong Impact on Post-Training Quantization for MNIST Dataset?

Driven by the need for the compression of weights in neural networks (NNs), which is especially beneficial for edge devices with a constrained resource, and by the need to utilize the simplest possible quantization model, in this paper, we study the performance of three-bit post-training uniform quantization. The goal is to put various choices of the key parameter of the quantizer in question (support region threshold) in one place and provide a detailed overview of this choice’s impact on the performance of post-training quantization for the MNIST dataset. Specifically, we analyze whether it is possible to preserve the accuracy of the two NN models (MLP and CNN) to a great extent with the very simple three-bit uniform quantizer, regardless of the choice of the key parameter. Moreover, our goal is to answer the question of whether it is of the utmost importance in post-training three-bit uniform quantization, as it is in quantization, to determine the optimal support region threshold value of the quantizer to achieve some predefined accuracy of the quantized neural network (QNN). The results show that the choice of the support region threshold value of the three-bit uniform quantizer does not have such a strong impact on the accuracy of the QNNs, which is not the case with two-bit uniform post-training quantization, when applied in MLP for the same classification task. Accordingly, one can anticipate that due to this special property, the post-training quantization model in question can be greatly exploited.     

新型单模大模场红外硫系玻璃光子晶体光纤设计研究

大模场光子晶体光纤在高功率激光传输、光纤放大器、光纤激光器中的广泛应用, 使其受到研究者的广泛关注.硫系玻璃在红外波段(1–20μm)具有优良透过性能, 且具有折射率高(2.0–3.5)、声子能量低(小于350 cm^-1)、 组分可调等特性, 成为制备红外光纤的理想材料. 本文设计一种基于Ge₂₀Sb₁₅Se₆₅硫系玻璃基质的新型单模传输、低损耗、超大模场面积光子晶体光纤结构, 经理论验证其在λ =10.6 μm处基模限制损耗远低于0.1 dB/m, 高阶限制模损耗大于2 dB/m, 模场面积约为13333 μm². 关键词： 硫系玻璃 大模场面积 红外光子晶体光纤 结构设计     

Airy加速光束的局域空间频率及其在光束设计中的应用

Airy加速光束是近年来备受关注的一种新型无衍射光束. 它所具有的自由加速、无衍射及自恢复特性使其在光学微操纵、非线性光学、 电子加速等诸多领域显示出重要的应用价值. 因此, 如何方便高效地生成加速光束成为近年来的一个热点研究内容. 本文对Airy加速光束复振幅分布的空间振荡特性进行了分析, 建立了利用局域空间频率描述其加速特性的理论. 提出了利用零点坐标计算加速光束局域空间频率的方法, 通过非线性拟合给出了可以精确描述Airy光束局域空间频率的解析公式; 确定了加速光束的局域空间频率函数与加速轨迹之间的定量关系, 给出了由给定加速轨迹计算相应的局域空间频率以及加速光束的纯相位函数的一种简单计算方法. 将上述分析结果用于设计产生具有给定加速轨迹的加速光束所需的相位函数, 成功求出了能够产生圆弧形加速轨迹的新型加速光束的纯相位函数的解析表达式. 基于该相位函数设计的纯相位衍射光学元件的模拟衍射结果证明了上述方法的可行性. 关键词： 衍射光学 加速光束 局域空间频率     

两相流有限穿越可视图演化动力学研究

针对小管径两相流流动特性, 全新优化设计弧形对壁式电导传感器. 通过动态实验在获取传感器测量信号的基础上, 采用有限穿越可视图理论构建对应于不同流型的两相流复杂网络. 通过分析发现, 有限穿越可视图网络异速生长指数和网络平均度值的联合分布可实现对小管径两相流的流型辨识; 有限穿越可视图度分布曲线峰值可有效刻画与泡径大小分布相关的流动物理结构细节特征; 网络平均度值可表征流动结构的宏观特性; 网络异速生长指数对流体动力学复杂性十分敏感, 可揭示不同流型演化过程中的细节演化动力学特性. 两相流测量信号的有限穿越可视图分析为揭示两相流流型的形成及演化动力学机理提供了新途径. 关键词： 两相流 复杂网络 有限穿越可视图 网络异速生长指数     

磁刺激穴位复杂脑功能网络构建与分析

本文采用互信息方法对磁刺激内关穴过程中的脑电信 号进行了两两通道间非线性时域关联特性分析, 构建了不同频率刺激前、刺激中、刺激后的脑功能网络, 并基于复杂网络理论对脑功能网络的特征进行了深入研究. 结果表明, 磁刺激频率为3 Hz 时, 大脑功能网络的平均度、平均聚类系数和全局效率与刺激前相比均有显著升高, 平均路径长度显著降低, 并且相应脑功能网络的"小世界"属性有所增强, 信息在大脑各区域间的传递更加高效. 本研究首次开展了磁刺激穴位复杂脑功能网络的构建与分析, 为探索磁刺激穴位对大脑神经调节的作用和机理提供新思路和新方法. 关键词： 复杂网络 磁刺激 脑功能网络 互信息