Evaluating intelligence: a computational semiotics perspective
Ricardo Gudwin · 2002
This paper discusses methods for evaluating the intelligence of intelligent systems by means of computational semiotics. Instead of looking at the system as a black box and testing its behavior, the process described focuses on architectural details of structures, organizations, processes and algorithms used in the construction of the intelligent system, evaluating the impact of using these elements in the overall intelligent behavior exhibited by the system. It proposes "insider" metrics that, coupled to "outsider" metrics, will be important for the determination of general metrics for intelligence in intelligent systems.