Deep-Machine-Learning-In-Architecture-Application

Explores the application potentials with single-objective optimization, multiple objective optimizations, supervised learning, unsupervised learning and deep learning in architectural fields. This study focuses on developing an optimization workflow for global structural form-finding and geometry morphologies.

Full Content Please See the PDF Documentation

Software Tool & Plugin:

• Rhino 3D
• Grasshopper 3D
• Python
• Anaconda
• MOpossum
• Goat
• Octopus
• Galapagos

Algorithm

• Single-Objective Optimization
• Multi-Objective Optimization
• Supervised Learning
• Unsupervised Learning
• Deep Learning
• Generative Deep Learning

Library:

• Tensorflow
• Pandas
• Scikit-learn

[Goal] To Find Most Optimized Shape(s) For Free-Form Concrete Shell Using Novel Strategies

Single-Objective Optimization [SOO]

Multi-Objective Optimization [MOO]

Unsupervised Learning

Supervised Learning

Deep Learning

Generative Deep Learning: Style Transfer for Shell Substructure

Generative Deep Learning: Brick Facade Patterning