A programmable neural virtual machine based on a fast store-erase learning rule
A programmable neural virtual machine based on a fast store-erase learning rule
复制标题
DOI：
10.1016/j.neunet.2019.07.017
复制
发表时间：
2019-11-01
期刊：
影响因子：
7.8
通讯作者：
Reggia, James A.
中科院分区：
文献类型：
--
作者：
Katz, Garrett E.;Davis, Gregory P.;Reggia, James A.
我们提出了一种神经体系结构，它使用一种新的局部学习规则来表示和执行用传统汇编语言编写的任意符号程序。这种神经虚拟机(NVM)纯粹是神经计算的，但支持传统计算机体系结构的所有关键功能。与其他可编程神经网络不同，NVM使用快速非迭代局部学习、信息的分布式表示、与程序无关的电路、巡回吸引子动力学以及用于活动性和可塑性的乘法门控等原理。我们详细介绍了NVM，从理论上分析了它的性质，并进行了经验性的计算机实验，量化了它的性能，证明了它是有效的。(C)2019爱思唯尔有限公司。保留所有权利。
We present a neural architecture that uses a novel local learning rule to represent and execute arbitrary, symbolic programs written in a conventional assembly-like language. This Neural Virtual Machine (NVM) is purely neurocomputational but supports all of the key functionality of a traditional computer architecture. Unlike other programmable neural networks, the NVM uses principles such as fast non-iterative local learning, distributed representation of information, program-independent circuitry, itinerant attractor dynamics, and multiplicative gating for both activity and plasticity. We present the NVM in detail, theoretically analyze its properties, and conduct empirical computer experiments that quantify its performance and demonstrate that it works effectively. (C) 2019 Elsevier Ltd. All rights reserved.