Synthetic Architecture: AI-Generated Design Explorations

Description:

This workshop mainly focuses on the intersection of artificial intelligence and architecture, and covers exploring the creative potential of generative AI models to drive architectural design explorations. Throughout the workshop, participants will be introduced to a range of advanced generative diffusion models including Midjourney, Stable Diffusion, Runway ML, DALL-E, etc., and design workflows that incorporate these generative models. They will gain a comprehensive understanding of the capabilities and applications of these models specifically in the context of architectural design, and will learn to take control of these models of fuzzy logic. The workshop is structured as a combination of informative sessions, tutorials, design sessions and design reviews. Through a combination of informative sessions, demonstrations and practical exercises, participants will learn how to leverage these models to generate diverse design options, refine their ideas, and push the boundaries of architectural creativity. The workshop will encourage participants to think innovatively and venture beyond traditional design approaches. By the end of the workshop, participants can expect to have a comprehensive understanding of the potential and limitations of generative AI models in architectural design.

Learning Objectives :

With this workshop, participants will learn about design workflows that effectively integrate various generative AI models into architectural design processes. They will understand how to take control of these models, utilizing their fuzzy logic to enhance architectural creativity and exploration. Through design reviews and discussions, participants will develop critical thinking and evaluation skills. They will learn to assess the outputs of generative AI models in the context of architectural design, enabling them to make informed decisions and refine their design concepts effectively. The workshop will encourage participants to think innovatively and venture beyond traditional design approaches. Through practical exercises and design sessions, participants will learn how to leverage generative AI models to generate diverse design options, refine their ideas, and push the boundaries of architectural creativity.

Detail Schedule :

Day 1 (July 22, 17.00 – 21.00 UTC)

_Introduction to Generative AI (image generation) models 

_Midjourney prompt craft and workflows

_Out-painting and in-painting in DALL-E and post-processing

_Using ChatGPT for creating concepts

_Design session #1 Creating architectural concepts in Midjourney

Day 2 (July 23, 17.00 – 21.00 UTC)

_Advanced prompting in Midjourney

_Multi-prompts / Image prompts / Settings and Parameters•Settings and parameters (aspect ratio, seed, stylize, image weight, tile)
_Design session #2 Expanding architectural concepts in Midjourney

Day 3 (July 24, 17.00 – 21.00 UTC)
_Stable Diffusion settings and parameters
_Using ControlNET
_Other platforms and tools (clipdrop, RunwayML, dreamstudio …)
_Sources

_Design session #3 Expanding architectural concepts incorporating models

Day 4 (July 25, 17.00 – 21.00 UTC)_Design reviews of the final projects and Collective evaluation session

Prerequisites

*Required Skills: None

*Required software:

_Midjourney (minimum basic subscription required, standard subscription recommended)

_Stable Diffusion (SD WebUI Automatic 1111 or Google Colab Pro required)

_ControlNet

_DALLE-2 (optional)

_ChatGPT

_Google Colab

_Photoshop (optional)

*Required hardware: PC or Mac

Instructor :