[Free Live Streaming] TouchDesigner Vol.050&051 Intro to Deep Learning for Graphics Programmers – From Training to Inference

We welcome Scott Allen, a visual artist based in Tokyo!

Date（Note: This workshop will be held in two sessions.）

DAY1: 5th December 2021 15:00 – 18:00 UTC+9
DAY2: 12th December 2021 15:00 – 18:00 UTC+9

DAY1: Check the event time in your time zone
DAY2: Check the event time in your time zone

Language

日本語 / Japanese

Level

Intermediate
・Basic operation of TouchDesigner is required.
・This workshop is designed for beginners with or without knowledge of machine learning.

Watch this workshop on TDSW YouTube Live

DAY1: https://youtu.be/gkTm_g47gA8
DAY2: https://youtu.be/iDg1eI1r3Sc

Free to watch this workshop in real-time on the TDSW YouTube channel. With the $20 / month or higher tier, you can watch all the recordings of our regular online workshops including this one.

Timetable (subject to change) time in UTC+9

DAY1: 5th December 2021
15:00-18:00　Workshop (incl. 10 mins break x 2)
18:00-18:10   Update from TDSW
18:10-18:20　Close

DAY2: 12th December 2021
15:00-18:00　Workshop (incl. 10 mins break x 2)
18:00-18:10   Update from TDSW
18:10-18:20　Close

Contents (subject to change)

 With the development of deep learning, AI-based systems are now in operation everywhere. In graphics programming, we can use the results of various analyses from the camera, such as object recognition and pose estimation. By generating frames directly with generative models such as GAN and VAE, we can create expressions that are different from those of classical methods. In this lecture, as a first step to using deep learning, I will show you the process of training on an architecture designed by yourself using Google Colaboratory and inferring on TouchDesigner using a simple discriminator and generator.

Instructor

Scott Allen

Completed Institute of Advanced Media Arts and Sciences(IAMAS) in 2016, studied under Ryota Kuwakubo. Currently enrolled in the doctoral course at Keio University. Focusing on the projection media, he creates an installation aimed at deviating from recognition of currently fixed images using lasers, LEDs, and deep learning generated images. Also he performs in a style that creates images by physically interposing the mechanism of the projection device and adding modulation as an evangelist of “ZOGAKU”, which strictly means playing visuals like musical instruments.
He teaches at Keio University from 2018, Musashino Art University from 2019, Tama Art University and Advanced Media Arts and Sciences(IAMAS) from 2020, Meisei University from 2021 as a part-time lecturer.
Website
Twitter

Code of conduct

The TDSW community operates under the Berlin Code of Conduct. Event attendees should also follow this document.

https://berlincodeofconduct.org/ja/