A Rational Account of Categorization Based on Information Theory

arXiv cs.AI / 4/1/2026

💬 OpinionSignals & Early TrendsIdeas & Deep AnalysisModels & Research

共有:

Key Points

The paper proposes a new, information-theoretic rational theory of categorization to explain how humans form categories.
It tests the theory by comparing its predictions against results from classic categorization experiments by Hayes-Roth and Hayes-Roth (1977), Medin and Schaffer (1978), and Smith and Minda (1998).
The authors report that the proposed framework matches or outperforms several existing approaches, including independent cue/context models, Anderson’s rational model, and a hierarchical Dirichlet process model.
The work frames categorization as an information-efficiency/rational analysis problem and positions the model as a competitive alternative to prior probabilistic and rational-cognitive accounts.

Abstract

We present a new theory of categorization based on an information-theoretic rational analysis. To evaluate this theory, we investigate how well it can account for key findings from classic categorization experiments conducted by Hayes-Roth and Hayes-Roth (1977), Medin and Schaffer (1978), and Smith and Minda (1998). We find that it explains the human categorization behavior at least as well (or better) than the independent cue and context models (Medin & Schaffer, 1978), the rational model of categorization (Anderson, 1991), and a hierarchical Dirichlet process model (Griffiths et al., 2007).