Song Tao, Gong Faming, Liu Xiyu, Zhao Yuzhen, Zhang Xingyi
IEEE Trans Nanobioscience. 2016 Oct;15(7):666-673. doi: 10.1109/TNB.2016.2598879.
Spiking neural P systems (SN P systems) are a class of parallel and distributed spiking neural network models, which are inspired from the way biological neurons spiking and communicating by means of spikes. White hole rules, abstracted from the biological observation of neural information rejection, were recently introduced into SN P systems, by which a neuron consumes its complete contents when it fires. In this work, SN P systems with white hole neurons are proposed, in which each neuron has only white hole rules. The computational power of general and bounded SN P systems with white hole neurons are obtained. Specifically, it is achieved in a constructive way that i) general SN P systems (having both bounded and unbounded) white hole neurons are Turing universal as number generators; ii) bounded SN P systems with white hole neurons can only characterize semi-linear sets of numbers. These results show that "information storage capacity" of certain key neurons provides some "programming capacity" useful for SN P systems achieving a desired computation power.
脉冲神经P系统（SN P系统）是一类并行分布式脉冲神经网络模型，其灵感来源于生物神经元通过脉冲进行脉冲发放和通信的方式。白洞规则是从神经信息拒绝的生物学观察中抽象出来的，最近被引入到SN P系统中，通过该规则，神经元在激发时会消耗其全部内容。在这项工作中，提出了具有白洞神经元的SN P系统，其中每个神经元仅具有白洞规则。获得了具有白洞神经元的一般和有界SN P系统的计算能力。具体而言，以建设性的方式实现了：i）一般的SN P系统（具有有界和无界的）白洞神经元作为数字生成器是图灵通用的；ii）具有白洞神经元的有界SN P系统只能表征数字的半线性集。这些结果表明，某些关键神经元的“信息存储容量”为SN P系统实现所需的计算能力提供了一些有用的“编程能力”。
