Modeling Human-Like Color Naming Behavior in Context

2026-04-28

AI Summary

Existing neural agents in color-naming tasks produce lexicons that lack the convexity characteristic of human color categories, exhibiting systematic deviations from empirical human data. This work proposes a novel approach that integrates upsampling of rare color terms with multi-listener reinforcement learning within a referential game framework, combining supervised learning and interactive training. To explicitly evaluate structural alignment with human categorization, the method introduces a convexity metric to quantify geometric consistency of emergent categories. Experimental results demonstrate that the proposed approach significantly enhances the diversity and informativeness of the generated lexicons while improving their similarity to human color naming systems, achieving the closest match to human data in both convexity and semantic structure among current models.

Abstract

Modeling the emergence of human-like lexicons in computational systems has advanced through the use of interacting neural agents, which simulate both learning and communicative pressures. The NeLLCom-Lex framework (Zhang et al., 2025) allows neural agents to develop pragmatic color naming behavior and human-like lexicons through supervised learning (SL) from human data and reinforcement learning (RL) in referential games. Despite these successes, the lexicons that emerge diverge systematically from human color categories, producing highly non-convex regions in color space, which contrast with the convexity typical of human categories. To address this, we introduce two factors, upsampling rare color terms during SL and multi-listener RL interactions, and adopt a convexity measure to quantify geometric coherence. We find that upsampling improves lexical diversity and system-level informativeness of the color lexicon, while many-listener setups promote more convex color categories. The combination of moderate upsampling and multiple listeners produces lexicons most similar to human systems.

Problem

• color naming
• convexity
• human-like lexicons
• computational modeling
• color categories

Innovation

• color lexicon
• convexity
• upsampling
• multi-listener reinforcement learning
• pragmatic communication