Global Workspace Theory

From Wikiversity

Global Workspace Theory

In 1988 when Bernard J. Baars wrote A Cognitive Theory of Consciousness he proposed the framework of a theory that could be developed into an Artificial Consciousness. The concept consisted of a Blackboard connected to a centralized network that limited throughput and informed a population of independent agents, that competed for attention. Each agent would do a specific function, and thus the population as a whole would offer the functionality of a consciousness.

While this is a timely theory, and multi-agent systems are implementable, there is considerable discussion about the need for the BlackBoard function, because it takes on the role of a theater of the mind, and philosophically there are some problems with theeater theories.

The main problem with this theory, is that while multiagent systems have shown promise, they become decoherent very easily, and do not scale well. Another problem is that it uses an adhoc definition of consciousness, in that it doesn't really define consciousness so much as say that if we take consciousness apart and put the right pieces back together we will get consciousness again.

--Graeme E. Smith 03:03, 27 January 2009 (UTC)