Complex Cyber-Physical Networks： From Cybersecurity to Security Control

Complex cyber-physical network refers to a new generatio~ of complex networks whose normal functioning significantly relies on tight interactions between its physical and cyber compo- nents. Many modern critical infrastructures can be appropriately modelled as complex cyber-physical networks. Typical examples of such infrastructures are electrical power grids, WWW, public trans- portation systems, state financial networks, and the Interact. These critical facilities play important roles in ensuring the stability of society as well as the development of economy. Advances in informa- tion and communication technology open opportunities for malicious attackers to launch coordinated attacks on cyber-physical critical facilities in networked infrastructures from any Interact-accessible place. Cybersecurity of complex cyber-physical networks has emerged as a hot topic within this con- text. In practice, it is also very crucial to understand the interplay between the evolution of underlying network structures and the collective dynamics on these complex networks and consequently to design efficient security control strategies to protect the evolution of these networks. In this paper, cybersecu- rity of complex cyber-physical networks is first outlined and then some security enhancing techniques, with particular emphasis on safety communications, attack detection and fault-tolerant control, are suggested. Furthermore, a new class of efficient secure the achievement of desirable pinning synchronization control strategies are proposed for guaranteeing behaviors in complex cyber-physical networks against malicious attacks on nodes. The authors hope that this paper motivates to design enhanced security strategies for complex cyber-physical network systems, to realize resilient and secure critical infrastructures.

Identifying topologies and system parameters of uncertaintime-varying delayed complex networks

Node dynamics and network topologies play vital roles in determining the network features and network dynamical behaviors.Thus it is of great theoretical significance and practical value to recover the topology structures and system parameters of uncertain complex networks with available information. This paper presents an adaptive anticipatory synchronization-based approach to identify the unknown system parameters and network topological structures of uncertain time-varying delayed complex networks in the presence of noise. Moreover, during the identification process, our proposed scheme guarantees anticipatory synchronization between the uncertain drive and constructed auxiliary response network simultaneously. Particularly, our method can be extended to several special cases. Furthermore, numerical simulations are provided to verify the effectiveness and applicability of our method for reconstructing network topologies and node parameters. We hope our method can provide basic insight into future research on addressing reconstruction issues of uncertain realistic and large-scale complex networks.

A threshold effect of coupling delays on intra-layer synchronization in duplex networks

This paper investigates the impact of inter-layer coupling functions and intra-layer coupling delays on intra-layer synchronization regions and sychronizability. It is found that the inter-layer coupling functions have great influence on intra-layer synchronization regions, as well as on the intra-layer synchronizability. In particular, there exists an inter-layer coupling function such that the inter-layer coupling strength neither improves nor weakens the intra-layer synchronizability. Furthermore, no matter which one of three inter-layer coupling functions is chosen, a small intra-layer delay always keeps the intra-layer synchronized regions almost unchanged, implying that the small delay neither enhances nor suppresses the intra-layer synchronizability. At the same time the delay greatly frustrates the synchronizability in each layer when it is greater than some threshold. Our results may have potential applications for interconnected technological networks where communication delays are inevitably present.

Evolution and maintenance of cooperation via inheritance of neighborhood relationship

Cooperative behaviors are ubiquitous in nature and human society.It is very important to understand the internal mechanism of emergence and maintenance of cooperation.As we know now,the offsprings inherit not only the phenotype but also the neighborhood relationship of their parents.Some recent research results show that the interactions among individuals facilitate survival of cooperation through network reciprocity of clustering cooperators.This paper aims at introducing an inheritance mechanism of neighborhood relationship to explore the evolution of cooperation.In detail,a mathematical model is proposed to characterize the evolutionary process with the above inheritance mechanism.Theoretical analysis and numerical simulations indicate that high-level cooperation can emerge and be maintained for a wide variety of cost-to-benefit ratios,even if mutation happens during the evolving process.

Substrate concentration effect on gene expression in genetic circuits with additional positive feedback

In gene regulatory networks, gene regulation loops often occur with multiple positive feedback, multiple negative feedback and coupled positive and negative feedback forms. In above gene regulation loops, auto-activation loops are ubiquitous regulatory motifs. This paper aims to investigate a two-component dual-positive feedback genetic circuit, which consists of a double negative feedback circuit and an additional positive feedback loop(APFL). We study effect of substrate concentration on gene expression in the single and the networked systems with APFLs, respectively. We find that substrate concentration can tune stochastic switch behavior in the signal system and then we explore relationship of substrate concentration with positive feedback strength in aspect of stochastic switch behavior. Furthermore, we also discuss gene expression and stochastic switch behavior in the networked systems with APFLs. Based on analysis in the networked systems, we discover that genes express in some specific cells and do not express in the other cells when the expression achieves its steady state. These results can be used to well explain the character of regionalization in the expression of genes and the phenomenon of gene differentiation.

Finite-time adaptive consensus of a class of multi-agent systems

Multi-agent systems(MASs) are ubiquitous in natural and artificial systems. This paper aims to establish the finite-time adaptive consensus criterion for a class of MASs with nonlinear dynamics. Traditionally, the finite-time consensus criterion is often established based on the prior information on Lipschitz constants and the eigenvalues of Laplacian matrix. However, it is difficult to acquire the above prior information for most real-world engineering systems. To overcome the above difficulty, this paper develops the finite-time consensus criteria for a class of MASs with nonlinear dynamics via adaptive technique. In detail, we design the finite-time distributed node-based and edge-based adaptive consensus protocols for a class of MASs with fixed and switching topologies. Numerical simulations are also given to validate the proposed finite-time adaptive consensus criterion.

What makes for uniformity for non-contrastive self-supervised learning?

Recent advances in self-supervised learning(SSL) have made remarkable progress, especially for contrastive methods that target pulling two augmented views of one image together and pushing the views of all other images away. In this setting, negative pairs play a key role in avoiding collapsed representation. Recent studies, such as those on bootstrap your own latent(BYOL)and SimSiam, have surprisingly achieved a comparable performance even without contrasting negative samples. However, a basic theoretical issue for SSL arises: how can different SSL methods avoid collapsed representation, and is there a common design principle? In this study, we look deep into current non-contrastive SSL methods and analyze the key factors that avoid collapses.To achieve this goal, we present a new indicator of uniformity metric and study the local dynamics of the indicator to diagnose collapses in different scenarios. Moreover, we present some principles for choosing a good predictor, such that we can explicitly control the optimization process. Our theoretical analysis result is validated on some widely used benchmarks spanning differentscale datasets. We also compare recent SSL methods and analyze their commonalities in avoiding collapses and some ideas for future algorithm designs.

An overview on GNSS carrier-phase time transfer research

Global navigation satellite system(GNSS)carrier phase observations are two orders of higher accuracy than pseudo-range observations,and they are less affected by multipath besides.As a result,the time transfer accuracy can reach 0.1 ns,and the frequency transfer stability can reach 1×10^-15 with carrier phase(CP)method,therefore CP method is considered the most accurate and promising time transfer technology.The focus of this paper is to present a comprehensive summary of CP method,with specific attention directed toward day-boundary clock jump,ambiguity resolution(AR),multi-system time transfer and real-time time transfer.Day-boundary clock jump is essentially caused by pseudo-range noise.Several approaches were proposed to solve the problem,such as continuously processing strategy,sliding batch and bidirectional filtering methods which were compared in this study.Additionally,researches on AR in CP method were introduced.Many scholars attempted to fix the single-difference ambiguities to improve the time transfer result,however,owing to the uncalibrated phase delay(UPD)was not considered,the current studies on AR in CP method were still immature.Moreover,because four GNSS systems could be used for time-transfer currently,which was helpful to increase the accuracy and reliability,the researches on multi-system time transfer were reviewed.What’s more,real-time time transfer attracted more attention nowadays,the preliminary research results were presented.

面向近期量子处理器的量子神经网络研究进展

近年来,人工神经网络在各领域得到了广泛的应用,展现出强大的计算智能.与此同时,量子计算硬件也得到了飞速发展,近期量子处理器已具备较稳定的计算能力和抑制退相干能力,多家商用云量子计算机公司已能够为世界各地的学者们提供在线量子计算实验平台.在诸多技术领域的迅猛进展下,量子神经网络这一交叉领域也重获科学界的关注,涌现出了大量新的研究思路并取得了重要的实验进展.本文首先对早期量子神经网络的研究思路作简要阐述,然后对基于近期量子处理器的量子神经网络的数据编码、算法流程和量子电路设计等进行综述分析,最后总结量子神经网络领域的数个关键科学问题,对未来研究方向做出展望.

Preface

Complex networks are ubiquitous in our lives. Representative examples are the Internet, social networks, biological networks, E-commerce networks, electrical power grids, and larger-scale engineering systems. It is well known that the Internet has been a powerful engine for our societal evolution and technological innovation. Nowadays, network science and engineering faces fundamental challenges, such as understanding the complexity of various large-scale networks, developing new architectures and exploiting new substrates, and enabling new applications and new economics. To a better future, the complex networks in our lives will need to be better： more accessible, more reliable, more predictable, and more secure.