The Ramon Llull's Thinking Machine for Automated Ideation

2025-08-26

AI Summary

This study addresses the dual challenges of insufficient conceptual diversity and weak alignment with research frontiers in scientific idea generation. We propose a triaxial co-creative framework—inspired by Ramon Llull’s medieval combinatorial art—that organizes ideas along thematic, domain, and methodological axes, implementing an interpretable, interactive “Llullian Thought Machine.” Methodologically, we extract structured conceptual units from domain expert knowledge and conference paper corpora, then leverage large language models for controllable combinatorial prompting and idea synthesis. Our key contribution lies in formalizing Llull’s combinatorial logic as a modern computable model and enabling traceable, human-in-the-loop creative exploration. Experiments demonstrate that our generated research proposals significantly outperform baselines in conceptual diversity (+32%), alignment with cutting-edge literature (27% improvement in Top-10 citation coverage), and interpretability.

Technology Category

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

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Economics, Online Markets and Human Computation: Architectures and workflows that use LLMs for crowd work

Abstract

This paper revisits Ramon Llull's Ars combinatoria - a medieval framework for generating knowledge through symbolic recombination - as a conceptual foundation for building a modern Llull's thinking machine for research ideation. Our approach defines three compositional axes: Theme (e.g., efficiency, adaptivity), Domain (e.g., question answering, machine translation), and Method (e.g., adversarial training, linear attention). These elements represent high-level abstractions common in scientific work - motivations, problem settings, and technical approaches - and serve as building blocks for LLM-driven exploration. We mine elements from human experts or conference papers and show that prompting LLMs with curated combinations produces research ideas that are diverse, relevant, and grounded in current literature. This modern thinking machine offers a lightweight, interpretable tool for augmenting scientific creativity and suggests a path toward collaborative ideation between humans and AI.

Problem

• Automating research idea generation using combinatorial methods
• Mining compositional elements from experts and literature
• Enhancing scientific creativity through human-AI collaborative ideation

Innovation

• Combinatorial framework with three axes
• LLM-driven exploration of research ideas
• Lightweight interpretable tool for creativity