A failed attempt at machine learning for realtime sound design in Pure Data Vanilla.

I've previously shown artificial neurons (   • Pure Data Artificial Neuron Patch fro...  ) and neural networks (   • Pd Artificial Neural Network Patch fr...  ) in Pd, but here I attempted to take things to the next step and make a cybernetic system that demonstrates machine learning. It went good, not great.

This system has a "target" waveform (what we're trying to produce). The neuron takes in several different waveforms, combines them (with a nonlinearity), and then compares the result to the target waveform, and attempts to adjust accordingly.

While it fails to reproduce the waveform in most cases, the resulting audio of a poorlydesigned AI failing might still hold expressive possibilities.

0:00 Intro / Concept
1:35 SingleNeuron Patch Explanation
3:23 The "Learning" Part
5:46 A (Moderate) Success!
7:00 Trying with Multiple Inputs
10:07 Neural Network Failure
12:20 Closing Thoughts, Next Steps

LINKS:
Subscribe: https://www.youtube.com/user/SimonHut...
Official Website http://simonhutchinson.com/
bandcamp: https://simonhutchinson.bandcamp.com/...
facebook:   / simonhutchinsoncomposer  
Buy me a coffee: https://kofi.com/simonhutchinson
Buy my old gear (affiliate link): https://reverb.grsm.io/simon
Sign up for my mailing list: http://eepurl.com/hVs7bT

#puredata #machinelearning #artificialintelligence #sounddesign