Silicon-Based Imagination

Architectural Drawings with Machine Learning | Seminar - BSc./MSc.


Class

This seminar will explore the current possibilities of artificial intelligence models for generation, manipulation and analysis of architectural drawings. In the last half decade, AI generated architectural images went from the fringes of experimental architecture to a popular research topic around the world. This seminar will utilize mostly RunwayML, an interface to multiple types of AI types that is democratizing Machine Learning for artists and desginers.

The seminar will be composed of 3 assignements. Software tutorials will be taught online syncronously and asychronously and at the end of each task, a discussion will be held physically. Some topics and techniques being explored are: Style Transfer, Generative Adversarial Networks for image generation,  photogrammetry, image to text and text to image models, object detection and image segmentation.

 

Notes:

No prior software knowledge is required and no special hardware is necessary. The seminar will be taught in English.

Info

Modules:

  • MSc. Architektur und Städtebau:
    Datenräume
  • BSc. Architektur:
    Physical Computing

Lecturer:

  • Victor Sardenberg

Schedule:

  • First Class: Tuesday 13.10.2020
    10:00 - 12:00 online 

Further Information on stud.ip