Duane D. DeSieno

C++/Windows Development Expert

EDUCATION:

M.E. Industrial Engineering and Operations Research, Cornell University, 1973

B.S. Industrial Engineering and Operations Research, Cornell University, 1971

PATENTS:

Standalone word processing systems:

• U. S. Patent No. 4,264,217. Issued April 28, 1981.

• Canadian Patent No. 1,115,848. Issued January 5, 1982.

Graded Learning Device and Method Neural Network:

• U. S. Patent No. 4,933,871. Issued June 12, 1990.

Neural Network For Improved Classification of Patterns.

• U. S. Patent No. 5,371,809. Issued December 6, 1994.

PUBLICATIONS:

"Removing Random Noise From EKG Signals Using a Back Propagation Network", AEROSPACE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE (1987).

"Adding a Conscience to Competitive Learning", PROCEEDINGS OF THE IEEE CONFERENCE ON NEURAL NETWORKS, (1988).

"Trading Equity Index Futures with a Neural Network," THE JOURNAL OF PORTFOLIO MANAGEMENT, (Fall 1992)

"The Median Learning Algorithm for Robust Unsupervised Learning," PROCEEDINGS OF THE WORLD CONGRESS ON NEURAL NETWORKS, (1994).

PROFESSIONAL EXPERIENCE:

Owner of Logical Designs Consulting, Inc., a systems and software technical consulting firm (1985 - Present). Developer of THINKS Neural Networks for Windows simulator and development environment.

Instructor for the University of California, San Diego Extension (1989 - 1995). Teaching Applications of Neural Networks, and Advanced Applications of Neural Networks.

Namsys Inc., Toranto, Canada (1994). Developed hardware and software for currancy counting and recognition system using Fast Classification neural network.

HNC Inc., San Diego, CA (1987 - 1991). As a consultant, developed neural network software and applications including:

• Coded BPN, PNN, ELVQ, ART2, and GLN Neural Network Software.

• Microcoded vector math and image processing routines for the ANZAPlus and BALBOA i860 products.
• Heartbeat Demonstration using a neural network as a digital filter.
• Credit Scoring application for Avco.
• Handwritten Number Recognition technology for the IDEPT product and demonstration.
• Box Elimination and Segmentation algorithm for use in bank check reading and IDEPT product.
• Machine Print Recognition technology and demonstration.
• Competitive Learning model for use with foreign exchange data.
• Broom Balancer demo using the Graded Learning Network.
• Printed Circuit Board Identification demonstration using video camera frame as input.
• Lunar Lander simulation and demo using GLN for training.

Organizing committee for the First Annual Conference on Neural Networks, San Diego (1987). Member, International Neural Network Society (INNS), IEEE.

Teledyne Ryan Aeronautical, San Diego, CA (1985 - 1987). As a consultant, was in charge of software development for the Model 324 Remotely Piloted Vehicle, a real-time, multiprocessor system using the Intel 8086/80186 processors coded with PL/M and 8086 assembly language.

Hovik Medical Group, San Diego, CA (1984 - 1985). As Director of Engineering, was responsible for the development of the Cardio Companion and Olympulse EKG monitoring products, involving RF, Analog, Digital Signal Processing (using TMS320), Graphics and IBM PC interfacing technology.

Val Corporation, Arlington, VA (1984). As a consultant, designed motherboard, memory expansion, and Centronics interface for a Timex/Sinclair computer expansion product.

Pilara Data Systems Inc., San Diego, CA (1977 - 1984). As Vice President of Engineering and Manufacturing, was responsible for the complete design of the Pilara 1000 and Pilara 2000 word processing systems, including CPU board and high resolution bit mapped video graphics interface.

Tans World Airlines Inc., New York, N. Y. (1973 - 1977). As Senior Operations Research Analyst, developed computer program systems to decrease operating costs for the airline, including the Preferential Fueling System using linear programming to exploit fuel price differentials at different airports under allocation constraints to minimize total system fuel cost and the Charter Routing System using network theory to assign aircraft to charter flights to maximize the number of flights scheduled while minimizing the number of ferry flights.