Machine learning, signal and image processing

This page includes material of the "Machine learning, signal and image processing" introductory course in the 3rd-year option "Numérique et Vivant" at CentraleSupélec. This course is given with Clément Elvira.

Agenda

Common signal representations

• Jupyter notebook with solution (.html)
• Jupyter notebooks (.ipynb) and data

Use cases in audio and speech processing

• Part 1: Audio source separation based on the sparsity and spatial diversity of the source signals
  • Slides (.html)
  • Slides (.pdf)
• Part 2: Speech production and modeling
  • Slides (.html)
  • Slides (.pdf)

Linear predictive coding and the source-filter model of speech production

• Jupyter notebook (.html)
• Jupyter Notebook (.ipynb) and data

ML: modeling, inference, learning

• Slides (.html)
• Slides (.pdf)

Gaussian mixture model estimation with the EM algorithm

• Jupyter notebook (.html)
• Jupyter Notebook (.ipynb) and data

ML: supervised discriminative learning

• Slides (.html)
• Slides (.pdf)

Multinomial logistic regression

• Jupyter notebook (.html)
• Jupyter notebook (.ipynb)