A probabilistic account of exemplar and category generation.

摘要: People are capable of imagining and generating new category exemplars categories. This ability has not been addressed by previous models categorization, most which focus on classifying rather than them. We develop a formal account exemplar generation proposes that knowledge is represented probability distributions over categories, categories generated sampling from these distributions. evaluated in two pairs behavioral experiments. In the first pair experiments, participants were asked to generate novel category. second after observing several related The results suggest influenced both structural distributional properties observed we argue our data better explained alternative approaches.