时间因素对脉冲神经膜系统计算能力的影响

脉冲神经膜系统是一类分布式并行的神经网络计算模型.从模型的研究着眼点来看,脉冲神经膜系统是一类全新的脉冲神经网络模型,属于第三代神经网络计算模型的研究范畴.通过分析评述脉冲神经膜系统中时间因素对其计算能力影响,讨论了异步工作模式下,时间以及时间无关模式下的脉冲神经膜系统的计算能力不会减弱,即具有图灵通用性.这些结果可为脉冲神经膜系统进行模糊信息处理提供理论支持,也可为基于脉冲神经膜系统的离散时间建模方法提供可行性论据.最后,针对目前该领域研究的热点和难点,对该领域研究的公开问题和研究方向进行讨论.

Time features in the computational power of spiking neural P systems

Spiking neural P systems (shortly named SN P systems)were a class of parallel and distributed neural-like computing models.In terms of motivation of models,SN P systems were known as a new candidate of spiking neural networks.In this paper,results on time features to the computational power of the systems were discussed.It was found that the computational power of asynchronous,timed and time-free SN P systems would not decease,i.e.,they could achieve turing universality.These results provided a theoretical support to the fuzzy information representation and the way for discrete time-series modeling by SN P systems.Finally,some open problems and research directions in SN P systems were discussed for further research.