EMK

Deeprise

Deep Learning Based Building Form Design System

Data Visualization

Generative Design

Interface Design

Machine Learning

Carnegie Mellon University

Pittsburgh, PA, USA

2020

This research proposes a new way to model and design a high-rise building form by learning the morphological features that correspond to its own architectural design principles through deep neural networks. This research demonstrates that a deep generative model can grasp highly complex principles of architectural form and that design decisions for new forms can be processed based on deep learning of data. By further developing a generated form into schematic design, the system provides an example of a high-rise building design with a style distinctive to the deep generative model.

Collaborators:
Jinmo Rhee

◀︎ ▶︎
Deeprise
Deeprise

Carnegie Mellon University

Pittsburgh, PA, USA

2020

This research proposes a new way to model and design a high-rise building form by learning the morphological features that correspond to its own architectural design principles through deep neural networks. This research demonstrates that a deep generative model can grasp highly complex principles of architectural form and that design decisions for new forms can be processed based on deep learning of data. By further developing a generated form into schematic design, the system provides an example of a high-rise building design with a style distinctive to the deep generative model.

Collaborators:
Jinmo Rhee