Bayesian Inference

Thomas L. Griffiths, Joshua B. Tenenbaum, Charles Kemp

Research output: Chapter in Book/Report/Conference proceedingChapter

5 Scopus citations

Abstract

Inductive inferences that take us from observed data to underdetermined hypotheses are required to solve many cognitive problems, including learning categories, causal relationships, and languages. Bayesian inference provides a unifying framework for understanding how people make these inductive inferences, indicating how prior expectations should be combined with data. We introduce the Bayesian approach and discuss how it relates to other approaches such as the "heuristics and biases" research program. We then highlight some of the contributions that have been made by analyzing human cognition from the perspective of Bayesian inference, including connecting symbolic representations with statistical learning, identifying the inductive biases that guide human judgments, and forming connections to other disciplines.

Original languageEnglish (US)
Title of host publicationThe Oxford Handbook of Thinking and Reasoning
PublisherOxford University Press
ISBN (Electronic)9780199968718
ISBN (Print)9780199734689
DOIs
StatePublished - Nov 21 2012
Externally publishedYes

All Science Journal Classification (ASJC) codes

  • Psychology(all)

Keywords

  • Bayesian inference
  • Inductive inference
  • Learning
  • Rational models

Fingerprint Dive into the research topics of 'Bayesian Inference'. Together they form a unique fingerprint.

Cite this