Multiple heads outsmart one: A computational model for distributed decision making

AbstractDistributed cognition and decision making has been a topic of intense research in the recent years. In this paper, a computational model of distributed decision making using a community of predictive coding agents is developed. The agents are embodied multimodal entities and situated in a shared environment. They have different visibility of the environment due to unique sensory and generative models. We show that communication between agents helps each of them reach a shared decision in a way that cannot be reached by brain processes in a single agent. Using a simulated environment, we show that sensory limitations may lead to incorrect or delayed causal inferences giving rise to conflicts in the mind of a predictive coding agent, and communication helps to resolve such conflicts and overcome the limitations.

Return to previous page