Omnico
Applying Machine Learning to Architecture
Applying Machine Learning to Architecture
Applying Machine Learning to Architecture
Omnico is a case study into how machine learning may disrupt fields in ways unexplored. We utilized Nvidia's open-source StyleGAN to prototype how a fictional AI-led architecture firm, Omnico, might leverage globally crowdsourced civic data to foster community culture, create value, and change the way a modern city is designed.
Omnico is a case study into how machine learning may disrupt fields in ways unexplored. We utilized Nvidia's open-source StyleGAN to prototype how a fictional AI-led architecture firm, Omnico, might leverage globally crowdsourced civic data to foster community culture, create value, and change the way a modern city is designed.
Omnico is a case study into how machine learning may disrupt fields in ways unexplored. We utilized Nvidia's open-source StyleGAN to prototype how a fictional AI-led architecture firm, Omnico, might leverage globally crowdsourced civic data to foster community culture, create value, and change the way a modern city is designed.
Research, AI, ML, Prototyping, UI
January 2021 - March 2021
Art Center College of Design
Omnico is an AI-led architecture firm that uses machine learning to visualize urban futures using a globally crowd-sourced database. Anyone can contribute through omni’s co-op ownership model, omniGroup. From its continuously growing database, Omnico generates urbanization projects from curated datasets reflective of the local interactions and relationships between people and buildings within specific environments.
The prototype below was curated using traditional residential image data from Vietnam, Malaysia, and Cambodia with zero year parameters.
The Results of our image set from StyleGAN
Image / Data Collection
How do we preserve authenticity? How do you define it? We came to understand that no culture is without foreign influence, and that "authenticity" meant how people incorporate outside influence into their everyday lives - Including their buildings. We determined that the vast amount of variables that influence what something looks like should not be filtered, but included via image data and then indirectly associated with the details that we do know. So what do we know? The chart below depicts how we chose to define what data to collect, and how that data contributes to producing authentic pieces.
User Flow / Model
It was important to limit the extent to which a designer or contributor would have to manually input data in order to remain efficient, effective in their job, or require a low barrier of entry. The diagram below illustrates where in the system a user would interact with the model, and how their seemingly simple set of inputs are informative enough to produce effective results.
Omnico's Portfolio
How does Omnico present itself to the world, crowdsource data, and attract clients? We went on to design a a Machine Learning-led architecture portfolio, inspired by your favorite firms.
Additional Prototypes
