USA flag logo/image

An Official Website of the United States Government

LOW-COST MASSIVELY PARALLEL NEUROCOMPUTING FOR PATTERN RECOGNITION IN…

Award Information

Agency:
Department of Energy
Branch:
N/A
Award ID:
17496
Program Year/Program:
1992 / SBIR
Agency Tracking Number:
17496
Solicitation Year:
N/A
Solicitation Topic Code:
N/A
Solicitation Number:
N/A
Small Business Information
Computational Biosciences, Inc
Box 2090 Ann Arbor, MI 48106
View profile »
Woman-Owned: No
Minority-Owned: No
HUBZone-Owned: No
 
Phase 1
Fiscal Year: 1992
Title: LOW-COST MASSIVELY PARALLEL NEUROCOMPUTING FOR PATTERN RECOGNITION IN MACROMOLECULAR SEQUENCES
Agency: DOE
Contract: N/A
Award Amount: $50,000.00
 

Abstract:

CONNECTIONIST (NEURAL NETWORK) APPROACHES TO PATTERN RECOGNITION AND ANALYSIS HAVE ATTRACTED GREAT INTEREST RECENTLY BECAUSE OF THEIR FLEXIBILITY, ABILITY TO LEARN BY EXAMPLE, AND ABILITY TO "SELF-ORGANIZE" TO REVEAL HIDDEN PATTERNS AND RELATIONSHIPS IN THE INPUT DATA SET. NEURAL NETWORKS ARE INHERENTLY PARALLEL COMPUTATIONAL STRUCTURES AND THUS POTENTIALLY EXCELLENT CANDIDATES FOR IMPLEMENTATION ON MASSIVELY PARALLEL COMPUTERS. PARTICULARLY IN THE TRAINING STAGE, SERIAL IMPLEMENTATIONS OF CONNECTIONIST MODELS ARE OFTEN LIMITED WITH RESPECT TO EITHER THE SIZE OF A NETWORK THAT MAY PRACTICALLY BE SIMULATED OR THE NUMBER OF TRAINING TRIALS THAT ARE PRACTICAL. THE EXPLOSION IN THE SIZE OF MACROMOLECULAR SEQUENCE DATABASES (SUCH AS GENBANK) THAT IS ALREADY OCCURRING HAS CREATED A NEED FOR PATTERN ANALYSIS SOLUTIONS WITH COST-PERFORMANCE CHARACTERISTICS SUPERIOR TO THOSE CURRENTLY AVAILABLE. PHASE I OF THE PLANNED RESEARCH IMPLEMENTS EFFICIENT PARALLEL ALGORITHMS FOR ALL CRITICAL COMPONENTS OF A BASIC MULTILAYER, FEED-FORWARD NEURAL NETWORK MODEL. THESE INCLUDE A "LINEAR COMBINER" FOR THE MULTIPLICATION OF CONNECTION WEIGHT MATRICES WITH INPUT (OR ERROR) VECTORS, A SIGMOIDAL ACTIVATION FUNCTION TO INTRODUCE NONLINEARITY IN NEURON BEHAVIOR, AND A COMPLETE PARALLEL IMPLEMENTATION OF THE BACK-PROPAGATION (GENERALIZED DELTA RULE) TRAINING ALGORITHM. ONCE IMPLEMENTED, THE NETWORK IS BEING EVALUATED AGAINST ALTERNATIVE PARAMETERIZATIONS IN A PROTOTYPE DNA SEQUENCE PATTERN RECOGNITION TASK.

Principal Investigator:


0

Business Contact:

Small Business Information at Submission:

Computational Biosciences Inc
Po Box 2090 Ann Arbor, MI 48106

EIN/Tax ID:
DUNS: N/A
Number of Employees: N/A
Woman-Owned: No
Minority-Owned: No
HUBZone-Owned: No