9. Structure and Flexibility in Bayesian Models of Cognition
- Abstract
- Introduction
- Mathematical Background
- Inferring Clusters: How Are Observations Organized into Groups?
- Inferring Features: What Is a Perceptual Unit?
- Learning Functions: How Are Continuous Quantities Related?
- Conclusions
- Basic Bayes
- Parametric and Nonparametric
- Putting Them Together: Nonparametric Bayesian Models
Parametric and Nonparametric
Putting Them Together: Nonparametric Bayesian Models
Inferring Clusters: How Are Observations Organized into Groups?
- A Rational Model of Categorization
- Associative Learning
A Rational Model of Categorization
- Rescorla-Wagner model
- The linear-Gaussian model also has an interesting connection to classical learning theories such as the Rescorla-Wagner model (Rescorla and Wagner, 1972), which can be interpreted as assum- ing a Gaussian prior on w and carrying out Bayesian inference on w (Dayan, Kakade, Montague 2000; Kruschke, 2008).
- sensory preconditioning
- Despite the successes of the Rescorla-Wagner model and its probabilistic variants, they incorrectly predict that there should only be learning when the prediction error is nonzero, but people and animals can still learn in some cases.
- For example, in sensory preconditioning (Brogden, 1939), two cues (A and B) are presented together without an outcome; when A is subse- quently paired with an outcome, cue B acquires associative strength despite never being paired with the outcome.
Inferring Features: What Is a Perceptual Unit?
- A Rational Model of Feature Inference
- Choice Behavior
A Rational Model of Feature Inference
- Bayesian linear regression
- Basis Functions and Similarity Kernels
- Modeling Human Function Learning
Bayesian linear regression
Basis Functions and Similarity Kernels
Modeling Human Function Learning
- Concluding remarks
- Some Future Questions