TOSCA: A new biologically-inspired cognitive architecture

This talk will be a presentation of initial ideas we are bringing together to create a new comprehensive cognitive architecture that is inspired by what is known about brain computation. Currently, we are developing a framework for developing and experimenting with such architectures. In this talk I will discuss the motivation for creating such an architecture, how it will qualitatively differ from architectures such as Soar and ACT, but also how it will differ from purely connectionist approaches. I'll also discuss two of the central organizing principles of TOSCA, which have to do with bottom-up clustering of sequences of clusters (based on work by Rick Granger of Dartmouth) and reinforcement learning based on intrinsic reward. No results – just lots of smoke and mirrors.