Reading

  1. How can machine learning support people's existing creative practices? Expand people's creative capabilities?

Part of the attraction of using machine learning in a creative practice is well summed up by this quote:

Exploring and exploiting the ways that these algorithms can be misused.

I think there’s the draw of the unknown, in that the algorithms can come up with outputs that we may never have thought of, since they operate on different rules and “innate knowledge” than we, as people, do. And like in the example of the recipe names, these outputs can be delightful and funny.

In some other ways, people use machine learning to deliberately create constraints on their practice, challenging themselves to make more with a self-imposed limit. Without such constraints, sometimes there’s too many options, paralyzing creativity: subject, medium, purpose, size—endless choices. And sometimes it’s simply more interesting to create in a medium that wasn’t meant to be used in such a way (ex. a pen, paper, and a skim through some cookbooks would provide more than enough names for food dishes, yet none would have been nearly as interesting as what the machine provided).


Dream up and design the inputs and outputs of a real-time machine learning system for interaction and audio/visual performance. This could be an idea well beyond the scope of what you can do in a weekly exercise.

https://www.figma.com/embed?embed_host=notion&url=https%3A%2F%2Fwww.figma.com%2Ffile%2FsTez34aeLBtor2e8VEcCgQ%2FFlowchart-Template-(Community)%3Fnode-id%3D0%253A1


Coding Exercise:

After training

After training

Training automatic

Training automatic

Code OR fullscreen view

Process:

I wanted to make a model that would detect how many fingers someone was holding up and then have the corresponding number of popups, sort of like a self-inflicted annoyance.

I also wanted a model that would start collecting data based off the hand gestures themselves, and not saving with clicks.

I still ended up using a start button, and a kind of convoluted route by saving to different categories according to the height of each finger (ex. first category is when all 5 fingers are below the white line). I feel like there must be a more efficient way of doing it, but I’m not quite sure how.

Screen Shot 2022-10-06 at 3.39.08 PM.png

I was also able to save the model, but wasn’t quite sure how to use it in another sketch (tried and failed to understand this example) so I’ve just added in my folder (not in use).

Also, my popups failed don’t work—can’t figure it out (would probably be easier if I was able to have the model in a separate sketch).