Sunday, September 17, 2006

US Patent 7107252 - Pattern Recognition Using Nanostructures

Pattern recognition is a problem for which traditional computational systems are ill suited because a certain degree of flexibility is necessary to determine matching and non-matching patterns. Traditional systems built upon binary logic are too well defined to reliably deal with the inherent fuzziness in analyzing voice and image patterns. This patent is one of several patents from Alex Nugent an inventor and founder of Knowm Tech which is attempting to find solutions to pattern recognition by using nanostructures formed to create reconfigurable neural nets. It has been noted that pattern recognition is a key step to achieving strong AI. If Nugent's strategy is successful (or at least more successful than competing approaches) and can lead to strong AI this patent may well be one of the more valuable nanotechnology related patents. Claim 1 reads:

1. A pattern recognition system, comprising: a neural network based on nanotechnology, wherein said neural network comprises a plurality of nanoconnections disposed within a solution within a connection network of said plurality of nanoconnections; and a pattern input unit, which communicates with said neural network, wherein said neural network processes data input via said pattern input unit in order to recognize data patterns.