Cognitive Dynamic Networks

认知动态网络

• 批准号: RGPIN-2014-04464
• 负责人: Haykin, Simon
• 金额: $ 3.72万
• 依托单位: McMaster University
• 依托单位国家: 加拿大
• 项目类别: Discovery Grants Program - Individual
• 财政年份: 2018
• 资助国家: 加拿大
• 起止时间: 2018-01-01 至 2019-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=653269
• 关键词: Cognitive; Dynamic; Networks

SUMMARY**Research into Cognitive Dynamic Networks (CDN), inspired by the human brain, will build on the pioneering work done on Cognitive Dynamic Systems (CDS), research into which has been carried out by Simon Haykin, McMaster University, for about the last 10 years. In descriptive terms, the study of CDN embodies two basically different entities, but integrated together into one project: The Memory, and the Network. The Memory part consists of a layered structure, wherein each layer consists of a set of three cognits (standing for cognitive units), namely: perceptual cognit for perceiving the environment; executive cognit for action on the environment; and working-memory cognit that mediates between the perceptual and execute cognits and stabilizes them. As for the Network, its function is to integrate the sensors (responsible for generating the observables), the memory, and motor controllers under a single umbrella. The net result of this integration is a powerful CDN, which is stochastic self-organized, and complex; it is therefore capable of processing Big Data under the constraint of limited computational resources. Whereas the Memory looks to Cognitive Neuroscience for inspiration, the Network looks to the new fast-accelerating, broadly defined Network Science for mathematical and computational need**The CDN project embodies three tasks, summarized as follows: Task 1 involves implementation of the CDN, based on a new software module that can accommodate the processing of Big Data for relevant (i.e., valuable) information. Task 2 studies intrinsic properties of CDN, namely, Stability, Robustness for trustworthiness, and Emergence behaviour, as well as extrinsic properties of CDN, namely, Observability, Controllability, and Universality. Task 3 addresses three areas for which CDN could make a practical difference: Arctic studies, the Power grid, and National security across the Canada-United States border along the Great Lakes.

摘要**对受人脑启发的认知动态网络（CDN）的研究将基于对认知动态系统（CDS）的开创性工作，该研究由麦克马斯特大学西蒙·海金（Simon Haykin），麦克马斯特大学（McMaster University）进行了大约过去10年的研究。 用描述性的术语，CDN的研究体现了两个基本不同的实体，但共同整合到一个项目中：内存和网络。 记忆部分由一个分层结构组成，其中每一层由一组三个认知（代表认知单位）组成，即：感知认知以感知环境；执行对环境行动的认知；工作记忆的认知在感知和执行认知之间介导并稳定它们。 至于网络，其函数是集成传感器（负责生成可观察到的），内存和电动机控制器在单个伞下。 这种整合的最终结果是一个强大的CDN，它是随机的自组织和复杂的。因此，它能够在有限的计算资源的约束下处理大数据。 尽管记忆力为灵感的认知神经科学，但该网络将新的快速加速，广泛定义的网络科学用于数学和计算需求** ** CDN项目体现了三个任务，如下所示：任务1涉及CDN的实现，基于一个新的软件模块的实现，可以基于一个可容纳相关数据的新软件。 任务2研究CDN的固有特性，即，稳定性，可信赖性的鲁棒性和出现行为以及CDN的外在特性，即可观察性，可控性和普遍性。任务3介绍了CDN可以产生实际区别的三个领域：北极研究，电网和整个加拿大俄罗斯州沿着大湖区接壤的国家安全。