US5146420A
Expired / Lapsed
US
⏱️ 0.0 years remaining
Communicating adder tree system for neural array processor
Assignee: International Business Machines Corp
•
Filing: 1991-08-05
•
Granted: 1992-09-08
•
Expires: 2011-08-05
Abstract
Abstract The neural computing paradigm is characterized as a dynamic and highly computationally intensive system typically consisting of input weight multiplications, product summation, neural state calculations, and complete connectivity among the neurons. Herein is described neural network architecture for a Scalable Neural Array Process (SNAP) which uses a unique intercommunication scheme within an array structure that provides high performance for completely connected network models such as the Hopfield model. SNAP's packaging and expansion capabilities are addressed, demonstrating SNAP's scalability to larger networks. The array processor uses a special type of adder tree which computes in a first direction and communicates in a second direction. The adder tree is thus responsive to a compute state and a communication state. The adder tree has the ability to provide a first driver responsive to a compute state for communicating an adder output to a data path and a second driver responsive to the communication state for connecting the data path to the neuron inputs.
Inventors
- 👤 Stamatis Vassiliadis
- 👤 Gerald G. Pechanek
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-11-05 20:36
CPC Classifications
G; G06; G06N; G06N3/00; G06N3/02; G06N3/06; G06N3/063; G06N3/10
Claims Analysis
Total Claims: 7
Claim 1
Independent Claim
78 words
For use in a neural array processor including a connectivity network, an adder tree comprising: state means for defining a compute state and a communication state; adder means responsive to said compute state for combining at least two inputs to generate an adder output; first driver means responsive to said compute state for communicating said adder output to a data path; and second driver means responsive to said communication state for connecting said data path to said inputs.
Claim 2
Dependent Claim
78 words
For use in a neural array processor including a connectivity network, an adder tree comprising: state means for defining a compute state and a communication state; adder means responsive to said compute state for combining at least two inputs to generate an adder output; first driver means responsive to said compute state for communicating said adder output to a data path; and second driver means responsive to said communication state for connecting said data path to said inputs.
Claim 3
Dependent Claim
51 words
For use in the neural array processor of claim 1, an activity function generator comprising: function generator means responsive to said compute state and to said adder output for generating a neuron value; and third driver means responsive to said communication state for communicating said neuron value to said data path.
Claim 4
Dependent Claim
193 words
A computation and communication means for use in a neural array processor including an adder tree, said system comprising: state means for defining a compute state and a communication state; adder means responsive to said compute state for combining at least two inputs to generate an adder output; first driver means responsive to said compute state for communicating said adder output to a data path; second driver means responsive to said communication sate for connecting said data path to said inputs; an activity function generator, said activity function generator further comprising means responsive to said compute state and to said adder output for generating a neuron value, and third driver means responsive to said communication state for communication said neuron value to said data path; a multiplier, said multiplier further comprising first operand means for storing a first operand and second operand means for storing a second operand, said multiplier responsive to said compute state and to said first and second operands for generating an output for communication to said adder means; and communication means responsive to said communication state for communicating said output on said data path to said second operand means.
Claim 5
Dependent Claim
193 words
A computation and communication means for use in a neural array processor including an adder tree, said system comprising: state means for defining a compute state and a communication state; adder means responsive to said compute state for combining at least two inputs to generate an adder output; first driver means responsive to said compute state for communicating said adder output to a data path; second driver means responsive to said communication sate for connecting said data path to said inputs; an activity function generator, said activity function generator further comprising means responsive to said compute state and to said adder output for generating a neuron value, and third driver means responsive to said communication state for communication said neuron value to said data path; a multiplier, said multiplier further comprising first operand means for storing a first operand and second operand means for storing a second operand, said multiplier responsive to said compute state and to said first and second operands for generating an output for communication to said adder means; and communication means responsive to said communication state for communicating said output on said data path to said second operand means.
Claim 6
Dependent Claim
149 words
A computation and communication system for use in a neural array processor including an adder tree, said system comprising: state means for defining a compute state and a communication state; adder means responsive to said compute state for combining at least two inputs to generate an adder output; first driver means responsive to said compute state for communicating said adder output to a data path; second driver means responsive to said communication state for connecting said data path to said inputs; a multiplier, said multiplier further comprising first operand means for storing a first operand and second operand means for storing a second operand, said multiplier responsive to said compute state and to said first and second operands for generating an output for communication to said adder means; and communication means responsive to said communication state for communicating said output on said data path to said second operand means.
Claim 7
Dependent Claim
149 words
A computation and communication system for use in a neural array processor including an adder tree, said system comprising: state means for defining a compute state and a communication state; adder means responsive to said compute state for combining at least two inputs to generate an adder output; first driver means responsive to said compute state for communicating said adder output to a data path; second driver means responsive to said communication state for connecting said data path to said inputs; a multiplier, said multiplier further comprising first operand means for storing a first operand and second operand means for storing a second operand, said multiplier responsive to said compute state and to said first and second operands for generating an output for communication to said adder means; and communication means responsive to said communication state for communicating said output on said data path to said second operand means.
References Cited (Backward Citations)
US4891782A
Cited
US5040134A
Cited
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 1995-09-08 | 1996-03-09 | 1996-09-08 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 1999-09-08 | 2000-03-09 | 2000-09-08 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2003-09-08 | 2004-03-09 | 2004-09-08 | $7700.00 | Window Closed / Lapsed |