拆解复杂互联系统中的信息流
Dismantling the information flow in complex interconnected systems
摘要：
微观结构损伤，如神经系统的病变或城市交通网络的中断，可能会损害系统功能的关键动力学，如电化学信号或人流，或任何其他类型的信息交换，分别在更大的拓扑尺度。损坏通常通过逐步移除组件或连接来建模，因此，系统的鲁棒性根据其结构分裂成断开的子系统的速度来评估。然而，这种方法未能捕捉到损害如何阻碍跨尺度的信息传播，因为即使在没有碎片的情况下，系统功能也会退化-例如，病态但结构完整的人脑在这里，我们探讨了复杂网络上的动态过程的损坏的响应，研究如何影响这样的信息流。我们发现，删除网络连接的节点中心可能有微不足道的影响，挑战传统的假设，结构指标本身就足以获得有关复杂系统如何运作的见解。使用一个破坏性的协议明确占流动动力学，我们分析合成和经验的系统，从生物基础设施的，并表明，它是可能的驱动系统的功能碎片完全结构解体之前。
英文摘要：
Microscopic structural damage, such as lesions in neural systems or disruptions in urban transportation networks, can impair the dynamics crucial for systems' functionality, such as electrochemical signals or human flows, or any other type of information exchange, respectively, at larger topological scales. Damage is usually modeled by progressive removal of components or connections and, consequently, systems' robustness is assessed in terms of how fast their structure fragments into disconnected sub-systems. Yet, this approach fails to capture how damage hinders the propagation of information across scales, since system function can be degraded even in absence of fragmentation -- e.g., pathological yet structurally integrated human brain. Here, we probe the response to damage of dynamical processes on the top of complex networks, to study how such an information flow is affected. We find that removal of nodes central for network connectivity might have insignificant effects, challenging the traditional assumption that structural metrics alone are sufficient to gain insights about how complex systems operate. Using a damaging protocol explicitly accounting for flow dynamics, we analyze synthetic and empirical systems, from biological to infrastructural ones, and show that it is possible to drive the system towards functional fragmentation before full structural disintegration.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
