ACT-RΦ: A cognitive architecture with physiology and affect

Abstract

Abstract Current computational cognitive architectures lack a comprehensive representation of human physiology. Connecting a more realistic model of physiology to a cognitive architecture allows one to develop more realistic computational process models that include an account for the effect of physiology on cognition. It is also useful to include some theory of affect and emotion to act as a functional layer between some representations of physiology and cognition. We chose to extend the ACT-R (adaptive control of thought – rational) cognitive architecture with physiology and affect to make the ACT-RΦ hybrid architecture. We connected the HumMod physiological simulation system with the ACT-R architecture; ACT-RΦ has a module that communicates between ACT-R and HumMod. A new affect system was also developed using an existing theoretical framework to serve as a functional layer between some of the physiological and cognitive representations. With this new architecture one can model a wide range of human behavior (e.g., the effects of homeostasis on cognition) in a more tractable fashion.