Natural Language Understanding System Based on Optoelectronic Neural Networks

NASA has stated a need for a technology which can provide efficient human interaction with intelligent systems. Specifically, natural language understanding is key to this interaction. In recognition of this need, Physical Optics Corporation (POC) proposes to develop a new real-time natural language understanding system, utilizing a new pattern recognition architecture recently developed at POC. This new architecture is based on the unique integration of an optical neural network architecture based on N4 interconnections, a high efficiency holographic memory matrix, and new fast electronic processing algorithms for speech recognition and statistical data analysis. The resulting system will be capable of highly robust speaker invariant natural language understanding. This system will drastically improve the efficiency of human interaction with intelligent systems in manned space missions. This will be achieved through the utilization of intriguing optical processing properties, such as large data storage and inherent high-speed parallel processing and interconnectivity.