Building synthetic brains capable of human level discovery and invention...

National Instrument’s LabVIEW distinguishes itself from traditional programming languages as an easy-to-use graphical or “G-programming” environment that includes all the functionality required for data acquisition, data analysis, presentation of results, and control of external devices. During project development, the user drags icons representing various native functionalities into a so-called “block diagram” and then interconnects them into a circuit that is effectively the G-program’s source code. Once such a graphical algorithm, what National Instruments calls a "virtual instrument" (VI), has been debugged, it may be compiled into a free-standing executable.

With support from the U.S. Air Force Research Laboratory, IEI has spent the last four years wrapping its patented neural network paradigms, such as STANNOs, Creativity Machines, Group Membership Filters, and SuperNets into LabVIEW VIs. As a result, the company now possesses a phenomenal capacity to rapidly prototype highly adaptive and creative control systems for robots. Extensive sensor integration may be achieved via the STANNO VIs, which may accept many millions of inputs from myriad sensors to interpret the system’s environment and challenges. Creativity Machines, formed from their constituent STANNOs, can then process such fused sensor data, enabling them to spontaneously invent the necessary robotic responses. These behavioral decision patterns may then be fed through a variety of VI-based electrical interfaces and abstraction layers to a wide variety of robotic actuators.

Using IEI’s patented toolbox for LabVIEW, these robotic systems are capable of knitting together vast swarms of STANNOs into SuperNets that may learn through successive self-experimentation to master both themselves and their environments. In fact, such a self-connected tabula rasa system, built over a period of just three days, was able to achieve autonomous docking and rendezvous of space vehicles within NASA simulators using a combination of IEI advanced machine vision and tabula rasa VIs. Even more recently, IEI has pioneered tabula rasa learning techniques in LabVIEW for NASA’s off-world robotics program.

Having built an extensive collection of camera VIs for a wide range of analog, USB, and IP cameras, IEI has fast-forwarded by decades in the area of machine vision applications, creating revolutions in the way target classification and change detection algorithms are conceived and implemented. They effectively wire themselves up to form the equivalent of the visual pathways of the brain. These systems are therefore capable of much more than simply recognizing what they are looking at. They actually understand what they see.

Ironically, IEI's G-Programming Toolbox has proven very useful in both prototyping and refining advanced Creativity Machine applications (i.e., conceptual discovery) unrelated to measurement and control problems. Furthermore, the IEI G-Programming Toolbox has shown exceptional results in building very extensive hierarchical cascades that are purely connectionist and adaptive systems that far excel over recently popularized Bayesian (i.e., lookup table) techniques. Currently, both Creativity Machine and hierarchical cascade function has been greatly accelerated, 10X or more, by placing most of the computational burden on highly parallel Graphical Processing Units (GPU).

Whereas the IEI G-Programming Toolbox is not available to the public, it is accessible to our growing number of licensed IEI development partners and carve-out ventures. Therefore, if you are excited about the possibility of futuristic products and services that effectively wire themselves together, contact our visionary business manager at IEI.