Artificial Intelligence | Online Workshop | English | North-South Americas
Description:
This workshop aims to teach legitimate means of three dimensionalizing the outputs of 2d GANs. The first half of the workshop will be focused on training neural networks on google colab. The students will be expected to train their own StyleGAN Model from a selection of data sets. The first day will also involve a lecture component in which the incorporation of machine learning into architectural workflows will be discussed. During the second session we will use the output of the neural network as an input into a parameterized 3 dimensionalization process. This script will be built in Houdini live with the students attending. This specific technique is one derived from two semesters of research at Texas A&M University and uses VDBs (Volumetric, Dynamic Grid that shares several characteristics with B+trees) in houdini in conjunction with the latent walk produced by the trained stylegan model to create 3 dimensional outputs. These outputs will then be textured using style transfer. The final objs with the style transferred texture will be rendered in keyshot. The ultimate goal of the workshop is to democratize the knowledge of neural networks as well as to pioneer the incorporation of these into digital workflows.
Key Words:
GAN,Language,VDB,Houdini
Required Skills:
Beginner Houdini, Basic Google colab, Beginner Rhino
Required Software:
Houdini, Rhino, Keyshot, Photoshop, Google Drive
Required Hardware:
Computer (capable of rendering 16 GB RAM)
Maximum number of participating students:
50
Schedule:
Jun 28 - Jun 29
-
Day 1 / Jun 28
-
Day 2 / Jun 29
Instructors:
-
Shane Bugni Texas A&M University,TAShane Bugni is an architecture student at Texas A&M University. He graduated from the Career and Technical Education Center in Frisco, Texas, in 2018. He has been teaching weekly online software tutorials since his sophomore year. Beginning in the fall of 2020 he has Co-taught ARCH 317, a research based design studio focusing on AI and robotic fabrication with a variety of collaborators including Joris Putteneers, Jean Jaminet, and Gabe Esquivel. He was a team lead in Gabriel Esquivel's T4T studio with Barry Wark. His work has been featured in Axiom, Ohio State's fulfilled exhibition, and the Venice Biennale. -
Gabriel Esquivel Texas A&M University,Associate ProfessorGabriel Esquivel was born and educated as an architect in Mexico City with a degree from the National University and received his Master’s Degree in Architecture from The Ohio State University. Gabriel is the director of the T4T Lab at Texas A&M University since 2010 where he examines the integration of digital technology to exchange architectural information and its connection to contemporary theory. He is also the director of the AI Advanced Research Lab.