Research
Pseudo Random
Automating 100,000 unique and workable architecture plans for AI trainingUniversity of Kansas
2025
In collaboration with:
Xu Han, PhD
Assistant Professor
School of Engineering
University of Kansas
This project focuses on the generation of 100,000 unique architectural plans for single housing units by using DrawScript and Shape Machine. Each plan is crafted to meet stringent quality checks. For the first attempt, here shows 9,346 unique plane, audience may access all 9,346 plans via the links below:
0-999
1000-1999
2000-2999
3000-3999
4000-4999
5000-5999
6000-6999
7000-7999
8000-8999
9000-9345
The goal is to create a dataset of high-quality, varied designs that can be used to train an advanced AI model focused on residential architecture. By exposing the AI to such a wide range of plans, the model learns to recognize design patterns, optimize space usage, and adapt to various environmental and aesthetic requirements. This large and diverse set of housing plans allows the AI to better understand architectural principles such as flow, light, and circulation, as well as factors like safety and efficiency. Over time, the model will be capable of generating new, optimized housing designs that meet specific criteria, or offering design suggestions for improvements. This project not only advances AI in architecture but also aims to improve the efficiency and innovation of future residential design and construction, contributing to Ai-aided design systems.
0-999
1000-1999
2000-2999
3000-3999
4000-4999
5000-5999
6000-6999
7000-7999
8000-8999
9000-9345
The goal is to create a dataset of high-quality, varied designs that can be used to train an advanced AI model focused on residential architecture. By exposing the AI to such a wide range of plans, the model learns to recognize design patterns, optimize space usage, and adapt to various environmental and aesthetic requirements. This large and diverse set of housing plans allows the AI to better understand architectural principles such as flow, light, and circulation, as well as factors like safety and efficiency. Over time, the model will be capable of generating new, optimized housing designs that meet specific criteria, or offering design suggestions for improvements. This project not only advances AI in architecture but also aims to improve the efficiency and innovation of future residential design and construction, contributing to Ai-aided design systems.
