感知网络中的最优信道效率
Optimal Channel Efficiency in a Sensory Network
摘要:
我们表明,在木之内-Copelli模型的雪崩寿命的分布的熵总是达到最大值的动态范围。这是值得注意的和不平凡的,因为虽然动态范围是感觉系统对刺激的敏感性的平衡平均度量,但弛豫时间的熵是一个纯粹的动态量,与刺激速率无关,可以解释为网络的效率,被视为一个通信信道。新发现的优化发生在我们测试的所有拓扑结构中,即使雪崩寿命的分布本身不是幂律,并且雪崩的大小分布的熵没有随之最大化,这强烈表明允许相互作用的神经元的状态的适当时间匹配的动力学规则是在信息处理中实现良好性能的关键,而不是增加可用单元的数量。
英文摘要:
We show that the entropy of the distribution of avalanche lifetimes in the Kinouchi-Copelli model always achieves a maximum jointly with the dynamic range. This is noteworthy and nontrivial because while the dynamic range is an equilibrium average measure of the sensibility of a sensory system to a stimulus, the entropy of relaxation times is a purely dynamical quantity, independent of the stimulus rate, that can be interpreted as the efficiency of the network seen as a communication channel. The newly found optimization occurs for all topologies we tested, even when the distribution of avalanche lifetimes itself is not a power-law and when the entropy of the size distribution of avalanches is not concomitantly maximized, strongly suggesting that dynamical rules allowing a proper temporal matching of the states of the interacting neurons is the key for achieving good performance in information processing, rather than increasing the number of available units.
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。