Data processing circuits in a neural network for processing first data stored in local register simultaneous with second data from a memory

• US 5,426,757 A
• Filed: 09/01/1992
• Issued: 06/20/1995
• Est. Priority Date: 01/24/1990
• Status: Expired due to Term

• Current Assignee: Renesas Electronics Corporation
• Original Assignee: Hitachi, Ltd.
• Inventors: Kimura, Katsutaka, Watanabe, Takao, Kawajiri, Yoshiki, Itoh, Kiyoo
• Primary Examiner(s): ELMORE, REBA I
• Application Number:
  US07/938,755
• Time in Patent Office: 1,022 Days
• Field of Search: 364/200 MS, 364/900 MS, 395/400 MS, 395/425 MS
• US Class Current: 711/101
• CPC Class Codes: G06N 3/063 using electronic means

Abstract

Herein disclosed is a data processing system having a memory packaged therein for realizing a large-scale and high-speed parallel distributed processing and, especially, a data processing system for the neural network processing. The neural network processing system according to the present invention comprises: a memory circuit for storing neuron output values, connection weights, the desired values of outputs, and data necessary for learning; an input/output circuit for writing or reading data in or out of said memory circuit; a processing circuit for performing a processing for determining the neuron outputs such as the product, sum and nonlinear conversion of the data stored in said memory circuit, a comparison of the output value and its desired value, and a processing necessary for learning; and a control circuit for controlling the operations of said memory circuit, said input/output circuit and said processing circuit. The processing circuit is constructed to include at least one of an adder, a multiplier, a nonlinear transfer function circuit and a comparator so that at least a portion of the processing necessary for determining the neutron output values such as the product or sum may be accomplished in parallel. Moreover, these circuits are shared among a plurality of neutrons and are operated in a time sharing manner to determine the plural neuron output values. Still moreover, the aforementioned comparator compares the neuron output value determined and the desired value of the output in parallel.

Claims

1. In a neural network, a data processing system constructed on a semiconductor chip, the data processing system comprising:

a memory having a plurality of data lines, a plurality of word lines arranged to intersect with said plurality of data lines and a plurality of memory cells each of which being formed at an intersection of a one of said plurality of data lines and a one of said plurality of word lines;

a word line selector for selectively selecting a one of said word lines; and

,

a plurality of processing circuits collectively executing a parallel processing on a pair of data items with a first data item of the pair being different from the second data item of the pair, each of said plurality of processing circuits including a register for storing first data comprising the first data item which is output through said plurality of data lines to said plurality of processing circuits from first memory cells connected with a first word line of said plurality of word lines, wherein each of said plurality of processing circuits includes means for executing an arithmetic processing according to the stored first data in the register simultaneous with the second data item which is output through said plurality of data lines to said plurality of processing circuits from second memory cells connected with a second word line of said plurality of word lines.

2. The data processing system in a neural network according to the claim 1 further comprising:

a data write means for writing output data of said plurality of processing circuits into said memory.

3. The data processing system in a neural network according to the claim 1 further comprising:

a read/write means for i) reading data from said memory and delivering the read data to an output of said semiconductor chip and ii) writing data from outside of said semiconductor chip into said memory.