Elements for a general memory structure:: properties of recurrent neural networks used to form situation models
Elements for a general memory structure:: properties of recurrent neural networks used to form situation models
复制标题
DOI：
10.1007/s00422-008-0221-5
复制
发表时间：
2008-05-01
影响因子：
1.9
通讯作者：
Cruse, Holk
中科院分区：
文献类型：
--
作者：
Makarov, Valeri A.;Song, Yongli;Cruse, Holk
我们研究单个记忆项是如何存储的，假设环境中给定的情况可以以循环神经网络中的类突触耦合的形式表示。先前的数值研究表明，基于抑制或最大单元的特定架构可以成功学习静态或动态刺激（情况）。在这里，我们提供了有关学习过程收敛和网络对新刺激的响应的理论基础。我们表明，除了学习“简单”静态情况之外，nD 网络还可以学习和复制最多 n 个不同向量或帧的序列。我们发现了学习率的限制，并显示了在不同情况下训练期间发展的耦合矩阵，包括将网络扩展到非线性单元间耦合的情况。此外，我们表明特定的耦合矩阵为单元提供低通滤波器特性，从而将静态求和单元构建的网络与连续时间网络连接起来。我们还展示了在什么条件下可以使用此类网络通过模式完成来执行算术计算。
We study how individual memory items are stored assuming that situations given in the environment can be represented in the form of synaptic-like couplings in recurrent neural networks. Previous numerical investigations have shown that specific architectures based on suppression or max units can successfully learn static or dynamic stimuli (situations). Here we provide a theoretical basis concerning the learning process convergence and the network response to a novel stimulus. We show that, besides learning "simple" static situations, a nD network can learn and replicate a sequence of up to n different vectors or frames. We find limits on the learning rate and show coupling matrices developing during training in different cases including expansion of the network into the case of nonlinear interunit coupling. Furthermore, we show that a specific coupling matrix provides low-pass-filter properties to the units, thus connecting networks constructed by static summation units with continuous-time networks. We also show under which conditions such networks can be used to perform arithmetic calculations by means of pattern completion.