作者: Roger C. Schank , Gregg C. Collins , Lawrence E. Hunter
DOI: 10.1017/S0140525X00051578
关键词:
摘要: The inductive category formation framework, an influential set of theories learning in psychology and artificial intelligence, is deeply flawed. In this framework a necessary sufficient features taken to define category. Such definitions are not functionally justified, used by people, inducible system. Inductive depend on having access all only relevant features, which impossible but begs key question learning. crucial roles other cognitive processes (such as explanation credit assignment) ignored or oversimplified. Learning necessarily involves pragmatic considerations that can be handled complex processes.We provide alternative for according must based function. system invokes accomplish difficult tasks, makes inferences, analyses decides among potential specifies how when categories generated modified. We also examine the methodological underpinnings two approaches compare their motivations.