Artificial neuron and method of using same

• US 5,390,136 A
• Filed: 06/14/1993
• Issued: 02/14/1995
• Est. Priority Date: 06/14/1993
• Status: Expired due to Fees

• Current Assignee: Motorola, Inc. (Motorola Solutions, Inc.)
• Original Assignee: Motorola, Inc. (Motorola Solutions, Inc.)
• Inventors: Wang, Shay-Ping T.
• Primary Examiner(s): Malzahn, David H.
• Application Number:
  US08/076,602
• Time in Patent Office: 610 Days
• Field of Search: 364/754, 364/736, 395/27
• US Class Current: 708/620
• CPC Class Codes: G06F 2207/5523 Calculates a power, e.g. the square, of a number or a function, e.g. polynomials
• CPC Class Codes: G06F 7/552 Powers or roots , e.g. Pythagorean sums
• CPC Class Codes: G06N 3/063 using electronic means

Abstract

An artificial neuron, which may be implemented either in hardware or software, has only one significant processing element in the form of a multiplier. Inputs are first fed through gating functions to produce gated inputs. These gated inputs are then multiplied together to produce a product which is multiplied by a weight to produce the neuron output.

Claims

1. A neuron circuit comprising:

first latch means responsive to a plurality of inputs x_i for sequentially storing each of said inputs x_i and for sequentially providing each of said stored inputs x_i as an output thereof, wherein i is a positive integer;

counter/latch means responsive to the output of said first latch means and providing said stored inputs as an output thereof;

second latch means for providing an output;

a multiplier circuit having a first input responsive to the output of said counter/latch means and a second input responsive to the output of said second latch means, and generating a product;

a multiplexer for coupling either said product or a weight value W to an input of said second latch means; and

wherein said counter/latch means includes means, responsive to a plurality of values g_i, there being a value g_i corresponding to each value x_i, for counting g_i multiplication cycles of said multiplier circuit; and

wherein said multiplier includes means for multiplying said input x_i by said product during each of said multiplication cycles, and for multiplying said product by said weight value during one of said multiplication cycles;

whereby said neuron circuit generates an output of the form W x₁^g 1 x₂^g 2. . . x_n^g n.

2. The neuron circuit recited in claim 1, and further comprising an output latch having a first input responsive to said product and a second input responsive to said weight value W, and generating an output representing the output of said neuron circuit, whereby said weight value W can bypass said multiplier circuit.

3. The neuron circuit recited in claim 1, wherein said first latch means is responsive to an input clock signal.

4. The neuron circuit of recited in claim 3, wherein said input clock signal is generated by a clock generator when the value of g_i in said counter/latch means is zero.

5. The neuron circuit recited in claim 1, wherein said counter/latch means is responsive to an internal clock signal.

6. The neuron circuit recited in claim 1, wherein said second latch means is responsive to an internal clock signal (CLK).

7. The neuron circuit recited in claim 1, wherein each of said inputs x_i comprises a plurality of bits.