Learning Structured Preferences

Abstract

Learning the preferences of other people is crucial for predicting future behavior. Both children and adults make inferences about others’ preferences from sparse data and in situations where the preferences have complex internal structures. We present a computational model of learning structured preferences which integrates Bayesian inference and utility-based models of preference from economics. We experimentally test this model with adult participants, and compare the model to alternative heuristic models.


Back to Table of Contents