Collective intelligence systems are intended for aggregation of agents’ opinions on something (mostly on the probability of some event) in such a way that the aggregated opinion approximates the opinion of a hypothetical omniscient agent who owns all the information. We outline a CI theory that incorporates empirical and theoretical considerations: we rethink and combine topics from various fields such as probability interpretations, probabilistic logic, decision theory, statistics, and algorithmic game theory. We also describe how CI for medical studies should be designed.

https://youtu.be/fwEQ-VtYo9g