Technology Background:

 

If our brain operated like computers it would store the image of every person’s face we ever met along with the person's name. Then when we meet someone, our brain would search through all these images to find a similar match and recognize who it is. Eventually, as we meet more people, it would take too long to recognize anybody - Obviously, our brain does not work that way.

 

This  immense gap between computer learning and human learning of patterns has been the focus of research by the founder for over 25 years. His research has focused on gaining insights from neural models of pattern recognition and incorporating them in computer algorithms. The PME is modeled after a generally accepted theory that our brain represents (encodes) sensory patterns in the form of neural oscillations. That is, states of neural activity that repeat in a cycle dependent on the specific sensory input pattern.Similarly, PME represents patterns as a dynamic repeating cycle of addresses in computer memory that is generated simply by random sampling the pattern. Different patterns create different cycles and similar patterns produce similar cycles. As a result similarity and clustering is inherent in the representation. PME’s unique implementation of this concept is the underlying reason for its simplicity and speed.  

 

Early versions of PME have resulted in 6 patents, two successful previous startups, and thousands of successful installations. The current version implements the similarity recognition with additional patents granted and pending.