Genesis is a desktop application created to support the PhD thesis in Computer Science and Music I completed at Durham University between 2009-2013. The key focus areas are Generative AI and Machine Creativity in the context of musical performance and composition.
It leverages Generative Processes including Genetic Algorithms, Fractals, Symbolic AI and Stochasticity to help musicians explore soundscapes in real-time.
The thesis predates the rapid growth and increased availability of LLMs, so it represents a point in time when Generative AI was accomplished by using 'traditional' algorithmic approaches.
The outcome of Genesis supported the notion I presented in my thesis to propose:
Generative processes are most successful when we partner as 'human and machine', creating new and innovative outputs together.
A point of view I still hold to this day, LLMs or not...
You can see the full thesis here.
I share the code here in the hopes that it brings value to the Generative AI community.
To use Genesis, you will need to download SuperCollider, and get an instance of it running. Then, simply load the Genesis file from this repo and you should be away.
-
I expect that you may need to install some classes that do not come as part of the standard SuperCollider install
-
If so, SuperCollider provides comprehensive error messaging which should help you identify what's missing
-
Use the guide here to help installation
-
-
SuperCollider has an incredible community which should also be able to help with general SuperCollider questions
-
There are certainly opportunities to optimise and improve the code. I welcome contributors to enhance and evolve Genesis!
Feel free to reach out to me on LinkedIn if you would like to learn more.