PSYC 532 Lecture Notes - Lecture 11: Posterior Probability, Bayes Estimator, Causal Inference
23 Jan 2018
School
Department
Course
Professor
Document Summary
Recent history of prominent cognitive modeling paradigms: rules & frames, neural networks, bayes. Conditioning since 2005, covers all the basic topics in cognition (from cognitioning to inductive inference, vision, language acquisition) all these techniques overlap historicaly over the years. Characteristics of bayesian models: rational updating of beliefs given new evidence, choose best hypothesis to explain observed data, integrates prior knowledge & new learning, exists between symbolic structures & subsymbolic networks, marr"s computational level. Statistically selects best structure to fit data. Hypotheses for coughing: we are meeting lot hypothesis: Previous causal inference models has the highest value prior (common) and likelihood. Baysien method someone and they are coughing a the most likely is 1 because it for prior. Graph 0: the background is necessary for the effect but not causal (not sufficient) Graph 1: the background is necessary for the effect, but also sufficient. The baysien model is one with the highest evidence in reasoning in the human brain.
