The Wisdom of Crowds: Diverse Intelligences and Collective Predictions and Classifications

Explore the foundations of prediction and classification in this comprehensive lecture by Scott Page from the University of Michigan. Delve into generated signals, categories, and dimensional reductions as models for humans, fauna, statistical predictions, and AI. Examine key theorems including the Condorcet Jury Theorem, Diversity Prediction Theorem, Bias-Variance Decomposition Theorem, Category Prediction Theorem, Two-Population Theorem, and the Hong Page Diversity-Accuracy Classification Theorem. Conclude with two evolutionary models linking institutional and network structures to diversity levels and collective accuracy. Recorded as part of IPAM's Mathematics of Intelligences Tutorials, this 76-minute talk offers valuable insights into the wisdom of crowds and diverse intelligences in collective predictions and classifications.