In the paper "Conscious Agent Networks", Chris Fields explains how conscious agents can be used to build memory and primitive kinds of learning, and how the pairs of these conscious agents can interact to satisfy a definition of a conscious agent.