Analysis of adaptation and environment

Designers often improve the performance of artificial agents by specializing them. We can make a rough, but useful distinction between specialization to a task and specialization to an environment. Specialization to an environment can be difficult to understand: it may be unclear on what properties of the environment the agent depends, or in what manner it depends on each individual property. In this paper, I discuss a method for analyzing specialization into a series of conditional optimizations: formal transformations which, given some constraint on the environment, map mechanisms to more efficient mechanisms with equivalent behavior. I apply the technique to the analysis of the vision and control systems of a working robot system in day to day use in our laboratory. The method is not intended as a general theory for automated synthesis of arbitrary specialized agents. Nonetheless, it can be used to perform post-hoc analysis of agents so as to make explicit the environment properties required by the agent and the computational value of each property. This post-hoc analysis helps explain performance in normal environments and predict performance in novel environments. In addition, the transformations brought out in the analysis of one system can be reused in the synthesis of future systems.