Understanding Collective Intelligence in Agent-Based Systems: an Information-Theoretic Approach to the Measurement and Comparison of Intelligence in Groups

This research aims to better understand the phenomenon of collective intelligence across the three cognitive systems: human, animal and machine. Major characteristics shaping the spread of intelligence in groups are identified and their influences on group performance are measured using an intelligence test. We address how much groups, implementing different communication and co-operation protocols, can outperform individuals by. Moreover, a new predictive model for the accuracy of artificial intelligence (AI) agents is presented. Finally, using notions from information theory, the studies of intelligence in AI are linked to other disciplines, notably business decision-making and management.