Most AI work now is being done with large Artificial Neural Networks and machine learning.

The System Modeling Institute is following a different path.  We are working with small spiking neural networks.  The thought is that it is easier to understand how nature “computes” using relatively small neural networks.  Our hypothesis is that useful neural network (NN) circuits exist in nature that are re-used over and over again.  Our goal is to discover these NN circuits.

We developed a tool that allows you to easily build and model spiking neural networks similar to how they operate in biological systems. It is useful as a “sandbox” to find and develop useful spiking neuron circuits that have useful behavior.   We have also used this tool to develop NN circuits that can control robots in the Webots platform.  (Webots is an open source tool that allows you to simulate robots in a virtual world).  We are currently exploring different NN circuit control ideas in order to generate interesting robot behavior.