A toolbox of methods for probabilistic inference

Abstract

We propose that probabilistic inference is supported by a mental toolbox that includes sampling and symmetry-based reasoning in addition to several other methods. To flesh out this claim we consider a spatial reasoning task and describe a number of different methods for solving the task. Several recent process-level accounts of probabilistic inference have focused on sampling, but we present an experiment that suggests that sampling alone does not adequately capture people's inferences about our task.


Back to Table of Contents