Bayesian Theories of Perception and Cognition
summary
Argues that perception uses probabilistic inference, tests it with psychophysics plus Bayesian model fitting to extract priors and likelihoods, and reproduces classic multisensory results (Ernst & Banks) where cues are weighted by reliability. Shows people learn simple stimulus priors very quickly—about 200 trials—producing hallucination-like false alarms and predictable estimation biases. Applies the same tools to clinical groups (clinical schizophrenia n≈22 patients; autistic traits in students), finding sharper sensory likelihoods with higher AQ and subtler effects in schizophrenia.
AI-written summary · about
chapters (9)
- 0:00Start
- 3:09The Uncertain History of the Bayesian Brain
- 4:38What is Bayes' theorem about?
- 8:13Cost functions - Bayesian Decision Theory
- 8:46This Talk
- 21:10Bayesian Cue Integration (7)
- 25:34A Bayesian theory of the Brain: Priors
- 56:36Computational Psychiatry
- 57:31Impaired integration of Priors in Schizophrenia
related talks
10mHallucinations: Harvesting Uncertainty
13mData-driven Sensory Intelligence
25mScaling Bayesian Experimentation
44mBayes' Theorem: The Formula That Changed the World
25mWhy Be Bayesian If I'm a Frequentist?
29mHow to use Data Science Superpowers in real life, a Bayesian perspective
42mBayes is BAE
40mProbabilistic Programming for Augmented Intelligence