#MachineLearningForArtists #CreativeCoding #CIIDSummerSchool2018 #SoloProject #GeneKongan #AndreasRefsgaard
Machine Learning for Artists was a 5-day hands-on workshops focusing on art & design applications of deep learning, including generative models, real-time tracking, and interactivity. We were introduced to diverse tools that allowed us to explore the subject without having to code a machine learning algorithm.
This was a small 2 day solo project as part of a workshop at the Copenhagen Institute of Interaction Design (CIID) Summer School 2018.
About the project
Conceptualised and executed in all of 48 hours, this project explores a fun way of using the Wekinator tool to train a computer to recognise an input and show an output associated with that input.
To make it fun, I came up with a storyline of two characters who were inseparable. I wanted to try pulling them apart being a notorious third friend.
It turned out pretty fun…
How I got there
In a nutshell
1. I trained the computer to recognise 3 different kinds of sounds with the help of the training software Wekinator, a microphone, and a voice input application.
2. I defined how my characters would move with each corresponding sound in Processing.
Detailed process
I trained Wekinator to identify three distinct sounds that I made into a microphone.
Used Processing to draw the characters and define how they’d move with the each sound.
Ran Wekinator to enable the input to control the output.
In this case, each sound (input) translated into the characters animating in some way (output), after it ran through the training algorithm (machine learning algorithm).
Result
Input: Tearing paper sound
Output: Characters separating like paper being torn apart
Input: “bzzzzz” like a bee sound
Output: Characters separating and flying around like bees
Input: “ciao” word sound
Output: Characters separating as one character disappears
Conclusion
Though a small project, I really went out of my comfort zone for this one. Deciding to do it on my own in a very short span of time, not being sure of the outcome, and not knowing the subject well, were a few situations I wouldn’t like to be in. This experience really helped me to focus on learning rather than the result and more, to believe in my ideas, however silly. This was also my first time playing writing code in Processing.
PS: This machine learning workshop was meant to be explored in artistic ways without having to dabble too much with the technicalities. It explored ways of using machine learning for more creative purposes.