Machine Learning techniques used in Artificial Intelligence systems became a buzz word that seems to solve everything without questioning how exactly it works.

Machine Learning algorithms form a complex system, like flocking birds or mycelium networks in the forest. The nature of complex systems allows it to behave in complex ways beyond our understanding, but the rules of the system are often very simple and understandable.

In this 2.5 day Lecture + Hands-on Workshop session, we dive into the basics of how Machine Learning functions, discuss how it can contribute within design and artistic projects as a tool, and learn how we can practically implement it through building a simple prototype.

Unlike Large Language Models (like Chat GPT), it is possible to train a small model with fewer data and less process time/energy using machine learning algorithms. This kinds of small models can be useful to analyze sensor data and create more intuitive interaction with the users. I will show some examples of textile sensors used in performance projects creating intuitive interactions between performers and sound systems.

In the hands-on workshop part, we use ml.lib (machine learning tool based on Gesture Recognition Toolkit, https://github.com/nickgillian/grt)) with Pure Data (https://puredata.info/) to implement a small machine learning training project. We can connect sensor inputs to this program and understand how machine learning tools can be used in analyzing sensors and how we can design intuitive interactions with it. The workshop is intended to support your semester project prototyping ideas. We can discuss how this technique can be applied in your project idea and how it can be implemented in your physical prototypes.

————————————

if you need something about me/bio

Mika Satomi is an artist, a designer and an educator working in the field of e-Textiles, Interaction Design and Physical Computing. Her work explores how we relate with technology and what we really want in them. She often collaborates with musicians and performers creating technology embedded costumes and interactive systems.

Since 2006 Mika has collaborated with Hannah Perner-Wilson, forming the art collective duo KOBAKANT creating artistic projects in the field of e-Textiles and Wearable Technology Art. She is a co-author of the e-Textile online database How To Get What You Want. Currently she lives and works in Berlin, Germany.

https://www.nerding.at/

https://www.kobakant.at/

http://howtogetwhatyouwant.at/