Behavioral Adaptive Autonomy. A milestone in the Alife route to AI?

Abstract

While central to robotics, biology and cognitive science, the concept of autonomy remains still difficult to make op-erative in the realm of Alife simulation models of cog-nitive agents. Its deep significance as a transition con-cept between life and cognition (a milestone on the Alife route to AI) remains obscured in the intricate relation be-tween metabolic/constructive processes and behavioral adap-tive processes in living systems. Within a naturalized and biologically inspired dynamical approach to cognition a defi-nition of behavioral adaptive autonomy is provided: homeo-static maintenance of essential variables under viability con-straints through self-modulating behavioral coupling with the environment, hierarchically decoupled from metabolic (con-structive) processes. This definition allows for a naturalized notion of behavioral adaptive functionality (that defines a proper level of modelling within Alife), structurally and in-teractively emergent: the mapping of the agent-environment system’s state space trajectories into the viability subspace of the essential variables of the organism.