IEI Sense Making Systems

Summary – Collectively, IEI’s patented neural network paradigms enable the construction of brain-like parallel, distributed algorithms capable of understanding, in the truest human sense, content of any kind arriving from the external environment. If that body of information relates to the national intelligence environment, our neural nets can effectively knit themselves into so-called “Supernets” that autonomously model the dynamics governing that world. By so doing, they go to heightened alert with newly arising scenarios, and automatically develop models of the principal entities therein and their interrelationships. Most importantly, they correlate such activities with any of the neural system’s analogous prior experience in a process equivalent to the way the human mind makes sense of the world. The IEI methodology is therefore a sine qua non for the national intelligence community as well as various corporate interests wanting a leg up on their competition.

Details – Whereas the public at large believes that creativity is confined to discovery, invention, and artistic endeavors, most of the mundane activities of the brain such as motor planning and the involuntary interpretation of sensory input are minor acts of creativity that are implemented through novel pattern generation followed by a tie-breaking process by other neural networks that shifts interpretation of raw sensory input in the direction of what is familiar, advantageous, or dangerous. In other words, understanding the world is actually the invention of significance to raw numerical patterns representing sensory inputs in a process tantamount to our Creativity Machine Paradigm.  In the world of intelligence gathering such raw inputs correspond to likewise numerical patterns from the intelligence environment that take the form of text, imagery, and various forms of acoustic and electromagnetic signals.

In 1996, the first generation of Supernets, having 100 million connection weights, self-organized itself into a vast compound neural net that was capable of integrating information from diverse intelligence streams.  In its “waking” state of operation, it could intercept both textual and graphical streams from multiple sources as sub-regions began to specialize themselves to specific genres of content. Later, in its “dreaming” stage, the Supernet could ceaselessly recall its direct experiences and interconnect them into deductive and inductive chains, coming to the self-realization that some of these revelations would be of particular interest to human analysts or authorities. In 1997, the same experiment was conducted, only now with 100 billion connection weights. With no customer in sight, it generated some very interesting philosophical revelations based upon the carefully selected closed world it was assigned to monitor!

The first important application of the IEI’s Creativity Machine Paradigm to sense making came in 2001 when a well-known intelligence agency approached the company to perform semantic data mining of the Internet in search of information related to certain terrorist activities. With the caveat that the system about to be donated to the US government was strictly in its beta test phase, the application performed admirably in what was to become a controversial and well-publicized pre-911 intelligence activity.

Recent activities in the last decade have included the building of sophisticated airport security applications that must integrate and make sense of signals arriving from sundry cameras, radar, and sensors. Autonomously building an immense model of airport status quo, it was able to isolate what didn’t make sense, to initiate a drill down to what mischief could follow.