公开(公告)号：US5956703A

公开(公告)日：1999-09-21

申请号：US508637

申请日：1995-07-28

申请人： Douglas D. Turner ,  Gabriela Breuer

发明人： Douglas D. Turner ,  Gabriela Breuer

IPC分类号： G06N3/063 ,  G06F15/18

CPC分类号： G06N3/063

摘要： A neural network IC 31 includes n dedicated processing elements (PEs) 62, an output register 66 for storing the PEs' outputs so that they are immediately accessible to all of the PEs, a number of output circuits 78 that are connected to selected PEs to provide binary outputs, and a timing circuit 74. Each of the PEs includes a weight memory 90 for storing input, output and bias weight arrays, a first in first out (FIFO) memory 88 for storing input data, a dot product circuit 92 and an activation circuit 94. The dot product circuit computes a dot product of the input weight array and the contents of the FIFO memory, a dot product of the output weight array and the contents of the output register, a dot product of the bias value and a constant, and sums the three results. The activation circuit maps the output of the dot product circuit through an activation function to produce the PE's output. The inclusion of a memory 90 that stores both input and output weight arrays in conjunction with the output register 66 allows the PEs to be configured to implement arbitrary feed-forward and recurrent neural network architectures.

摘要翻译： 神经网络IC31包括n个专用处理元件（PE）62，用于存储PE输出的输出寄存器66，使得它们能够被所有的PE容易地访问，多个输出电路78连接到所选择的PE 提供二进制输出和定时电路74.每个PE包括用于存储输入，输出和偏置权重阵列的权重存储器90，用于存储输入数据的先进先出（FIFO）存储器88，点积电路92和 激活电路94.点积电路计算输入权重阵列和FIFO存储器的内容的点积，输出权重数组的积和输出寄存器的内容，偏置值的积积和 一个常数，并总结三个结果。 激活电路通过激活功能映射点积电路的输出，以产生PE的输出。 包括存储器90，存储器90与输出寄存器66一起存储输入和输出权重数组，允许PE被配置成实现任意的前馈和循环神经网络架构。