Instructors: Dmitry Bogdanov, Rafael Ramirez, Pablo Alonso
Credits: 5 ECTS

This course overviews some of the common machine learning (deep learning) techniques used in sound and music audio analysis.The course is offered in 10 weeks, with 25 hours (2.5 h/week) of lectures with an overview of various topics, followed by a hands-on session and a weekly assignment. The evaluation of the students is based on the weekly assignments and participation in class.Materials for each topic, including slides and relevant code, will be made available before each class. All practical work will be conducted using Python, including libraries such as scikit-learn, Keras, PyTorch, PyTorch Lightning, and TorchAudio.

Topics covered

  • Introduction to ML and AI
  • Linear/logistic regression and gradient descent
  • Traditional machine learning algorithms (kNN, decision trees, etc.)
  • Artificial neural networks
  • CNNs, audio representations
  • Audio autoencoders
  • Metric learning for audio
  • Audio transformers
  • Audio embedding models and transfer learning applications
  • Text-audio models

Materials and References