Pragmatics Meets Culture: Culturally-adapted Artwork Description Generation and Evaluation

arXiv cs.AI / 4/6/2026

💬 OpinionIdeas & Deep AnalysisModels & Research

共有:

Key Points

The paper studies how language models perform in open-ended text generation when describing artworks for different cultural audiences with varying familiarity with embedded symbols and narratives.
It introduces a new task, “culturally-adapted artwork description generation,” to explicitly measure cultural familiarity and bias beyond decision-making benchmarks.
To evaluate cultural competence in this pragmatic setting, the authors propose an assessment framework using culturally grounded question answering.
Results show base models are only marginally adequate, while a “pragmatic speaker model” improves simulated listener comprehension by up to 8.2%.
A human evaluation corroborates the approach, finding that higher-pragmatic-competence outputs are rated as more helpful for comprehension by 8.0%.

Abstract

Language models are known to exhibit various forms of cultural bias in decision-making tasks, yet much less is known about their degree of cultural familiarity in open-ended text generation tasks. In this paper, we introduce the task of culturally-adapted art description generation, where models describe artworks for audiences from different cultural groups who vary in their familiarity with the cultural symbols and narratives embedded in the artwork. To evaluate cultural competence in this pragmatic generation task, we propose a framework based on culturally grounded question answering. We find that base models are only marginally adequate for this task, but, through a pragmatic speaker model, we can improve simulated listener comprehension by up to 8.2%. A human study further confirms that the model with higher pragmatic competence is rated as more helpful for comprehension by 8.0%.