DICE (Directions In Creative Entanglement）

 

🌟 Introduction

Project DICE is an AI-driven creative tool designed to help designers create every day, everywhere. This project integrates core concepts of design, inspiration, and creation, providing an efficient and innovative platform for designers to explore new creative realms and enhance design quality and efficiency. Utilizing multimodal large language models, DICE generates diverse and contextually relevant design inspirations, supporting structured creative exploration. One of our key exploration paths is the use of metaphors, a widely used method in design thinking.

We introduce Metaphor, a creativity support tool that connects targets (the products being designed) with sources (inspirational concepts), combining these two components to spark new design ideas. This tool is designed to help novice designers explore and develop metaphorical product concepts, offering richer inspiration and support throughout the creative process.

Metaphor: A creative thinking tool in the DICE system

🚀 Quick start

🛠️ setup Backend Service

In backend folder:

gunicorn productmeta_backend.asgi:application -k uvicorn.workers.UvicornWorker --workers 2 --bind 0.0.0.0:8000

🌐 Setup Frontend Service

In frontend folder:
1. Install dependencies using npm
npm install

2. Start the development server
npm run dev

👀 Metaphor Demo Presentation

📃 Paper

@unpublished{zhou2024Metaphor,  
	title = {Metaphor: Supporting Ideation in Metaphorical Product Design through Multimodal Large Models},  
	author = {Zhou, Qinyi and Deng, Jie and Liu, Yu and Wang, Yun and Xia, Yan and Ou, Yang and Lu, Zhicong and Li, Scarlett and Xu, Yingqing},  
	year = {2025},  
	note = {Conditionally accepted to CHI 2025},  
	selected = {true},  
	preview = {Metaphor.png},  
	pdf = {Metaphor_1210clear.pdf},  
	html = {/papers/paper1/}  
}

🤝 Contributing

Yu Liu, Yun Wang, Yan Xia, Yang Ou, Scarlett Li Collaborated with Tsinghua Future Lab and Hongkong City University

⚖️ Disclaimer

Due to the inherent limitations of large language models, issues such as hallucinations may occur.