Wisdom of crowds versus groupthink: learning in groups and in isolation (Q361811): Difference between revisions

The paper presents four classes of boundedly-rational strategies in multi-armed bandit problems and investigates if their asymptotic performance depends on whether the strategy is considered in isolation or in a network, or on whether the strategy is evaluated by itself or within a collection of communicating learners. Although it essentially is a theoretical paper, its findings can be useful for fields such as economics, psychology, computer science and philosophy, where various phenomena are modeled through such strategies.