Cooking Up Creativity: A Cognitively-Inspired Approach for Enhancing LLM Creativity through Structured Representations

2025-04-29

AI Summary

Large language models (LLMs) often lack genuine creativity, especially in generating semantically novel and feasible conceptual combinations. Method: This paper proposes a cognitive science–driven paradigm for structured creative generation: abstract concepts are modeled as executable, structured representations, and explicit recombination operations—inspired by human analogy and mental restructuring—are introduced to systematically explore creative possibilities in high-order semantic space, going beyond conventional lexical or syntactic perturbation. The approach integrates structured knowledge modeling, cognition-inspired recombination, domain-constrained reasoning, and LLM-augmented generation, instantiated in culinary recipe innovation. Results: Experiments demonstrate significant improvements over GPT-4o: +32.7% higher novelty scores in expert evaluation and +28.4% higher diversity according to automated metrics, while maintaining high feasibility and semantic coherence.

Technology Category

• Cognitive Modeling & Cognitive Systems: Computational Creativity
• Natural Language Processing: Generation
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information

Abstract

Large Language Models (LLMs) excel at countless tasks, yet struggle with creativity. In this paper, we introduce a novel approach that couples LLMs with structured representations and cognitively inspired manipulations to generate more creative and diverse ideas. Our notion of creativity goes beyond superficial token-level variations; rather, we explicitly recombine structured representations of existing ideas, allowing our algorithm to effectively explore the more abstract landscape of ideas. We demonstrate our approach in the culinary domain with DishCOVER, a model that generates creative recipes. Experiments comparing our model's results to those of GPT-4o show greater diversity. Domain expert evaluations reveal that our outputs, which are mostly coherent and feasible culinary creations, significantly surpass GPT-4o in terms of novelty, thus outperforming it in creative generation. We hope our work inspires further research into structured creativity in AI.

Problem

• Enhancing LLM creativity through structured representations
• Generating diverse and novel ideas beyond token-level variations
• Improving creative output in specific domains like culinary recipes

Innovation

• Couples LLMs with structured representations
• Recombines structured representations for creativity
• Demonstrates enhanced creativity in culinary domain