Self-repairing Hardware Paradigms based on Astrocyte-neuron Models

基于星形胶质细胞神经元模型的自修复硬件范例

• 批准号: EP/N007050/1
• 负责人: David Halliday
• 金额: $ 87.14万
• 依托单位: University of York
• 依托单位国家: 英国
• 项目类别: Research Grant
• 财政年份: 2015
• 资助国家: 英国
• 起止时间: 2015 至 无数据
• 项目状态: 已结题
• 来源: https://gtr.ukri.org/projects?ref=EP%2FN007050%2F1
• 关键词: Self; repairing; Hardware; Paradigms; based

The human brain is remarkable in its ability to self-repair, for example following stroke or injury. Such self-repair results from a range of distributed and fine-grained mechanisms which act in tandem to ensure that the neurones (the basic building blocks in the brain) continue to function in as close to a normal state as possible.In contrast modern electronic systems design typically relies on a single controller or processor, which has very limited self-repair capabilities. There is a pressing need to progress beyond current approaches and look for inspiration from biology to inform electronic systems design.Recent studies have highlighted that interactions between astrocytes (a type of glial cell) and neurones in the brain provide a distributed cellular level repair capability where faults that impede or stop neuronal firing can be repaired by a re-adjustment of the local weights of connections between neurones in the brain.This project aims to exploit these recent findings and develop a new generation of self-repairing algorithms by taking inspiration from these results to design a new generation of "astro-centric" algorithms. To achieve this we will include components representing both neurones and astrocytes in our electronic systems and model the interactions between these in such a way as to capture the distributed repair capabilities seen in the biological system.

人类大脑在自我修复能力方面非常出色，例如在中风或受伤后。这种自我修复是由一系列分布式和精细机制共同作用的结果，这些机制确保神经元（大脑的基本构建块）继续尽可能接近正常状态地发挥作用。相比之下，现代电子系统设计通常依赖于单个控制器或处理器，其自我修复能力非常有限。目前迫切需要超越现有的方法，从生物学中寻找灵感，为电子系统设计提供信息。（一种类型的神经胶质细胞）和神经元在大脑中提供了一种分布式细胞水平的修复能力，其中阻碍或停止神经元放电的故障可以通过重新修复来修复。调整大脑神经元之间连接的局部权重。本项目旨在利用这些最新发现，并从这些结果中获得灵感，设计新一代“以天文为中心”的算法，从而开发新一代的自我修复算法。为了实现这一目标，我们将在我们的电子系统中包括代表神经元和星形胶质细胞的组件，并以捕获生物系统中可见的分布式修复能力的方式对这些组件之间的相互作用进行建模。

项目成果

Fault-Tolerant Learning in Spiking Astrocyte-Neural Networks on FPGAs

• DOI: 10.1109/vlsid.2018.36
• 发表时间: 2018-03
• 期刊: 2018 31st International Conference on VLSI Design and 2018 17th International Conference on Embedded Systems (VLSID)
• 作者: Anju P. Johnson;Junxiu Liu;Alan G. Millard;Shvan Karim;A. Tyrrell;J. Harkin;J. Timmis;L. McDaid;D. Halliday

FPGA-based Fault-injection and Data Acquisition of Self-repairing Spiking Neural Network Hardware

• DOI: 10.1109/iscas.2018.8351512
• 发表时间: 2018-05
• 期刊: 2018 IEEE International Symposium on Circuits and Systems (ISCAS)
• 作者: Shvan Karim;J. Harkin;L. McDaid;B. Gardiner;Junxiu Liu;D. Halliday;A. Tyrrell;J. Timmis;Alan G. Millard;Anju P. Johnson

An FPGA-based hardware-efficient fault-tolerant astrocyte-neuron network

• DOI: 10.1109/ssci.2016.7850175
• 发表时间: 2016-09
• 期刊: 2016 IEEE Symposium Series on Computational Intelligence (SSCI)
• 作者: Anju P. Johnson;D. Halliday;Alan G. Millard;A. Tyrrell;J. Timmis;Junxiu Liu;J. Harkin;L. McDaid;Shvan Karim

Self-repairing mobile robotic car using astrocyte-neuron networks

• DOI: 10.1109/ijcnn.2016.7727359
• 发表时间: 2016-03
• 期刊: 2016 International Joint Conference on Neural Networks (IJCNN)
• 作者: Junxiu Liu;J. Harkin;L. McDaid;D. Halliday;A. Tyrrell;J. Timmis

Homeostatic fault tolerance in spiking neural networks utilizing dynamic partial reconfiguration of FPGAs

利用 FPGA 动态部分重配置的尖峰神经网络的稳态容错能力

• DOI: 10.1109/fpt.2017.8280139
• 发表时间: 2017
• 作者: Johnson A