Neural-Nets-Music-Hackspace

Materials for Neural Networks via FluCoMa workshop hosted by Music Hackspace

Dec. 11, 2022

Workshop Leader:

What to expect in this workshop?

In this workshop, Ted Moore from the FluCoMa project will guide you through an exploration of some of the creative possibilities available via Neural Networks with their Max Package. Basic experience of FluCoMa is advised before joining this workshop. For example, it is recommended that you have taken the free on-demand workshop Using Machine Learning Creatively via FluCoMa In Max.

What you'll learn

Train neural networks to perform musical tasks
Use training and testing data to validate trained models
Troubleshoot the training process by adjusting neural network parameters
Combine different types of input and output data

Outline

Training a Neural Network (refresher)

classification
- musical example: feed by Ted Moore
- how to train a classifier
live visuals (category switching, like lights but just with colors or whatever)
validation
overfitting vs underfitting

MLP Paramters

hiddenlayers
activation functions
learnrate
maxiter
batchsize
validation
momentum

Adjusting Parameters During Training

musical example: improv Ted Moore and Fabrizio Spera (first 2 min.)
training a regressor
regression with control data
- xy
- video pixels

Regression with Audio Analysis

creating the datasets
musical example: section from quartet by Ted Moore
predict fm parameters

Wavetable Autoencoder

wavetable interpolation with autoencoder

tedmoore/Neural-Nets-FluCoMa-Workshop