Research on the Architecture Design Based on Large Language Model
Xingchun Li, Min Li, Xiaokai Xia, Jingxiu Song, Xusong Bu · 2024
This paper presents an architecture design tool based on the large language model, it achieves intelligent architecture design by understanding user intentions, and generating and optimizing designs. The architecture design tool is responsible for receiving user design inputs and displaying design results, while the large language model handles the task of intelligent architecture design. To enhance the large language model’s understanding of architecture design, a prompt learning method based on the chain of thought is used, enabling the large language model to have more potent reasoning abilities and interpretability. The LangChain framework integrates the large language model with third-party tools to apply architecture design knowledge. The large language model empowers the architecture design tool, providing intelligent, efficient, and reliable design support.