Schematic Aspect for Autonomous Agent

Abstract

One serious issue for building an autonomous agent is that an agent does not understand the environment in which it is situated. To understand the world, the agent’s perception should be obtained directly from its own sensors instead of being provided indirectly by a human. Cohen suggests a physical schema as a conceptual primitive which enables an agent to percept a pattern directly. Cohen demonstrates that physical schema can be learned through the agent’s own sensorimotor activities. We propose negation as a conceptual primitive which enables an agent to recognize a schema itself. We also propose that emotive schema is necessary for building an autonomous agent as shown in the area of planning.