Mr. Ty Roachford, a Ph.D. student studying Theoretical and Computational Neuroscience in FAU’s Machine Perception and Cognitive Robotics Laboratory and the Rubin and Cindy Gruber Sandbox, will be presenting his research at the International Association for Perceptual Control Theory (IAPCT)
🔗 Time / Date / Info
🔗 Talk Description
Ensoul is a framework dedicated to studying, and generating models of, natural control systems. To this end, Ensoul pursues the development of what might be called “cognitive capacity tests”, which would be a set of tests used to understand the variable a system is controlling during a certain task, the variables a system is capable of controlling, and how one might modify the variable that a system is currently controlling. The Test for Controlled Variable (TCV) covers the first, while the second would be covered by the Test for Variable Controllability (TVC) and Test for Control Switch (TCS) respectively. The development of such tests is not a trivial matter, and to that end having a model system would facilitate the development and validation of such tests. In pursuit of this, Energy homeostatic loops, or enerstatic loops, were developed. Enerstatic loops are control systems that regulate their own causal dynamics using an “energy value” as a representation that captures the effects of those dynamics. Using this sort of computational model in conjunction with an evolutionary scheme, Ensoul hopes to create programs capable of evolving populations of natural control systems. This would provide a powerful tool for the development of cognitive capacity tests and in doing so, help further develop our approaches to studying living organisms as natural control systems.