This section summarizes the courses in machine learning applied to music computing given along the ATIAM Masters at IRCAM.

Evaluation

The evaluation of the ML course is defined by a group project. All details can be found on the corresponding projects page

Course summary

Introduction

  • Introduction to artificial intelligence
  • Properties of machine learning
  • Nearest-neighbors

Neural Networks

  • The artificial neuron
  • Neural networks
  • Architecture zoo

Support Vector Machines

  • Support Vector Machines
  • Properties of kernels

Unsupervised clustering

  • Clustering motivations
  • K-Means and k-medoids
  • Hierarchical clustering

Meta-heuristics

  • Genetic algorithms
  • Boosting

Probabilistic models

  • Probabilities and distributions
  • Belief networks

Bayesian inference

  • Bayesian learning
  • Undirected graphical models
  • Maximum likelihood

Gaussian Mixture Models

  • Expectation Maximization
  • Mixture models

Hidden Markov Models

  • Undirected graphical models
  • Hidden Markov models

Data complexity and GPU computing

  • Pitfalls of machine learning
  • Data complexity
  • GPU computing

Deep learning

  • Deep learning
  • Applications

Going further

  • Applications of ML
  • Current trends

Tutorials

The present tutorials covers coding exercices designed to implement the core notions seen in the machine learning lessons. Most techniques can be applied to any type of data from which sets of features can be computed. The exercices here target these techniques specifically applied to musical or audio data.

Get the baseline code for your language of choice :

Matlab zip file

Python zip file

Get the audio datasets from this link

Unzip files and place 00_Datasets along with other folders

Exercices

References