The SSM Sequence Models (SSMs) provide a mechanism for analyzing information and the relationships that may exist for that information in a much more computationally efficient manner than any current mechanisms in use today. In its simplest terms, the SSMs can provide a spell checker that can identify a misspelled word and provide the correct spelling of the actual intended word. In some of its more complex uses, the SSMs can provide voice recognition and speech synthesis, robotic learning using associative and auto associative memory, object recognition, Internet searching and categorization of information, and methods of recognizing, classifying, and analyzing biological sequences such as protein and DNA sequences–all with very high accuracy–to name a few. Indeed, SSMs may be used in any application that currently use Hidden Markov Models (HMMs), and will provide these systems with an increase in speed and accuracy, and a decrease in the computing power that is needed to accomplish the specific task. Further, unlike HMMs that often must be trained off line due to their computational complexity (particularly as the sequences involved become large), the SSMs can be trained in real time. Simply put, SSMs are much more efficient and effective than HMMs in performing all of the tasks for which HMMs are currently used, and therefore provide an elegant replacement.
Pattern or Sequence Recognition Applications Including, but Not Limited to, Voice Recognition, Objection Recognition, Computational Biology, Robotic Learning, Search, and Classification