![[2412.18899] GAI: Generative Agents for Innovation](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2412.18899] GAI: Generative Agents for Innovation
Summary
The paper explores GAI, a framework for generative agents that enhances collective reasoning to foster innovation, evaluated through a case study on Dyson's bladeless fan.
Why It Matters
This research highlights the potential of generative agents in simulating innovative processes, offering insights into how AI can contribute to creative problem-solving and development in various fields. Understanding these dynamics can inform future AI applications in innovation-driven industries.
Key Takeaways
- GAI framework enables collective reasoning among generative agents.
- Internal states in agents significantly enhance idea refinement and coherence.
- The framework was successfully tested using Dyson's bladeless fan as a case study.
- Models with heterogeneous agents outperformed simpler configurations.
- This research could influence future AI applications in innovation.
Computer Science > Artificial Intelligence arXiv:2412.18899 (cs) [Submitted on 25 Dec 2024 (v1), last revised 19 Feb 2026 (this version, v3)] Title:GAI: Generative Agents for Innovation Authors:Masahiro Sato View a PDF of the paper titled GAI: Generative Agents for Innovation, by Masahiro Sato View PDF HTML (experimental) Abstract:This study examines whether collective reasoning among generative agents can facilitate novel and coherent thinking that leads to innovation. To achieve this, it proposes GAI, a new LLM-empowered framework designed for reflection and interaction among multiple generative agents to replicate the process of innovation. The core of the GAI framework lies in an architecture that dynamically processes the internal states of agents and a dialogue scheme specifically tailored to facilitate analogy-driven innovation. The framework's functionality is evaluated using Dyson's invention of the bladeless fan as a case study, assessing the extent to which the core ideas of the innovation can be replicated through a set of fictional technical documents. The experimental results demonstrate that models with internal states significantly outperformed those without, achieving higher average scores and lower variance. Notably, the model with five heterogeneous agents equipped with internal states successfully replicated the key ideas underlying the Dyson's invention. This indicates that the internal state enables agents to refine their ideas, resulting in the const...