Culinary Crossroads: A RAG Framework for Enhancing Diversity in Cross-Cultural Recipe Adaptation

2025-07-29

AI Summary

In cross-cultural recipe adaptation, existing retrieval-augmented generation (RAG) methods exhibit cultural adaptability and relevance but suffer from severe output diversity deficiency—over-relying on local context fragments despite diverse input contexts, thus failing to accommodate heterogeneous user preferences. This paper proposes CARRIAGE, the first RAG framework that explicitly models and optimizes generative diversity. It introduces dual enhancements: (1) diversity-aware retrieval to broaden coverage over candidate recipes, and (2) a plug-and-play diversity control mechanism in context organization for multi-preference adaptive adaptation. Experiments demonstrate that CARRIAGE significantly improves output diversity while preserving adaptation quality, achieving Pareto-optimal trade-offs between diversity and fidelity. Our work establishes a novel paradigm for creative RAG tasks requiring both cultural sensitivity and user-aligned variability.

Technology Category

• Natural Language Processing: Generation
• Cognitive Modeling & Cognitive Systems: Computational Creativity
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Web Mining and Content Analysis: Mining multimedia, multimodal, multilingual, cross-lingual Web data

Abstract

In cross-cultural recipe adaptation, the goal is not only to ensure cultural appropriateness and retain the original dish's essence, but also to provide diverse options for various dietary needs and preferences. Retrieval Augmented Generation (RAG) is a promising approach, combining the retrieval of real recipes from the target cuisine for cultural adaptability with large language models (LLMs) for relevance. However, it remains unclear whether RAG can generate diverse adaptation results. Our analysis shows that RAG tends to overly rely on a limited portion of the context across generations, failing to produce diverse outputs even when provided with varied contextual inputs. This reveals a key limitation of RAG in creative tasks with multiple valid answers: it fails to leverage contextual diversity for generating varied responses. To address this issue, we propose CARRIAGE, a plug-and-play RAG framework for cross-cultural recipe adaptation that enhances diversity in both retrieval and context organization. To our knowledge, this is the first RAG framework that explicitly aims to generate highly diverse outputs to accommodate multiple user preferences. Our experiments show that CARRIAGE achieves Pareto efficiency in terms of diversity and quality of recipe adaptation compared to closed-book LLMs.

Problem

• Enhancing diversity in cross-cultural recipe adaptation
• Addressing RAG's over-reliance on limited context
• Generating varied outputs for multiple user preferences

Innovation

• CARRIAGE framework enhances RAG diversity
• Plug-and-play RAG for cross-cultural recipes
• Pareto efficiency in diversity and quality