Human collective intelligence is distinctly powerful. We collaborate with others to accomplish together what none of us could do on our own; we share the benefits of collaboration fairly and trust others to do the same. Even young children have the multi-agent common sense to understand, learn from, and collaborate with others in ways that are unparalleled in other animal species and are still lacking in our most sophisticated artificial intelligences. I will present a mathematical framework that combines hierarchical Bayesian models of learning with game-theoretic models of social interaction. This framework provides a computational basis for distinctly human aspects of multi-agent common sense which I test with behavioral experiments: inferring and forming joint intentions with others, reasoning about cooperation with common knowledge, and learning highly structured cultural knowledge from others given only sparse examples.
Copyright. All Rights Reserved.