Neural network In Architecture
Housing Constellations & Aggregate Worlds Speculations on Deep Architectures
Professor: Marcelo Spina, Michael Casey Rehm
Guest Instructors : Soomeen Hamn, Damjan Jovanovic
Assistant Teacher: Laure Michelon
Class: SCI-Arc - Summer 2020- MS. Architectural Technologies
Team: Yibo Qiao, Isabel Socorro
The project combines neural network with other learning intelligent agent algorithms to organize housing blocks, automating plan and massing production for modular housing structures.
This project also aims to investigate the relationship between architecture and its site and also the use of different technologies throughout the design process.
Dataset for PixtoPix
In this stage, the project is developing a program-based neural network that was trained to translate building footprints and massing slices into organizational diagram for programs.
Scripted Massing
Programmed Massing
Series of modular components
Diagrammatic plan and section show the interior space
