Paul S. Rosenbloom: “Implementing First-Order Variables in a Graphical Cognitive Architecture”

November 13, 2010 | Washington, DC

Speaker: Paul S. Rosenbloom
Host: 1st International Conference on Biologically Inspired Cognitive Architecture (BICA 2010)

Graphical cognitive architectures implement their functionality through localized message passing among computationally limited nodes. First-order variables, particularly universally quantified ones, while critical for some potential architectural mechanisms, can be quite difficult to implement in such architectures. A new implementation strategy based on message decomposition in graphical models is presented that yields tractability while preserving a priori symmetries in the graphs concerning how variables are represented and how symbols, probabilities and signals are processed.