|Realizing "Consciousness" In Software Agents|
| Realizing "Consciousness" In Software Agents describes the first design and implementation of Bernard Baars' global workspace theory. Global workspace theory is a leading psychological model of human consciousness. The "Conscious" Software Research Group at the University of Memphis has labeled agents which implement this theory as "conscious" software agents. As background material for the reader, this work also discusses agents, other existing cognitive architectures, and current software reuse methodology.
This dissertation describes in depth the "Conscious" Agent framework (ConAg), developed by this author. ConAg is a reusable software framework that carefully follows software reuse methodology. ConAg provides a solid foundation for building "conscious" software agents, and in particular, "consciousness" within these agents. A description of two agents built with ConAg are described, as well as the framework's structure. It is beyond this work's scope to address whether or not agents built with ConAg are sentient.
There are several motivators for this research. First, it is hypothesized that a global workspace gives a multi-agent system several advantages. For example, it offers individual agents in the system a means of recruiting the system's other agents to help solve novel and ambiguous problems. Also, it gives a method for attentional focus for the overall system. This provides a means for associative learning and metacognitive functions to take place.This dissertation gives an in depth discussion of the specific functions of the global workspace in ConAg. It is hoped that the software implementation advantages gained when using a global workspace are evident.
As a second motivation, this research hopes to provide new hypotheses about human consciousness for cognitive scientists and neuroscientists. As a cognitive science theory, Baars' theory does not contain many of the low-level design specifications necessary for a computer scientist's implementation of the theory. To implement the theory, these design decisions had to be made. Many of these decisions can be considered hypotheses on human consciousness. It is hoped that the members of the above disciplines will view these implementation decisions as springboards for the further study of consciousness.