Signal Processing for Artificial Intelligence study track

pictogramme cœurFor those who like


  • Mathematics applied to practical problems
  • Statistical learning
  • Signal processing


Upon completion of this track, students will have a broad and operational perspective of statistical learning and signal processing. They will understand the issues surrounding data processing and big data, the methodological foundations (statistics, optimization) and techniques for processing temporal data in particular (signal processing).

In practice

The teaching prioritizes rigorous lectures and practical work in realistic conditions.

Language of instruction: English

After the track

3rd year technological innovation at Télécom Paris

Master’s-Engineering Dual Degree

  • Automation and Signal and Image Processing (Univ. Paris-Saclay)
  • Data and Artificial Intelligence (IP Paris)
  • Data Science (IP Paris)
  • Mathematics, Vision, Learning (IP Paris/Univ.
  • Acoustics, signal processing and computer
    science applied to music (Sorbonne Univ.)
  • Bio-Imaging (Univ. Paris-Cité, Biomedical specification)


The track trains future engineers who will have a wide range of skills in the area of statistical learning (machine learning) and signal processing, which cover numerous fields of application: music and speech, biosignals, radio astronomy, transmission and compression of multimedia information, etc.


Yukun Liu

The track associates the knowledge from broad subjects, and these subjects are all explored step by step. For one example, the path of learning for machine learning is from Hilbert space to SVM, to perception, along to neural network. And this helps build solid foundations in the expertise. The track connects theory tightly with practice. Each course contains two or three practical works, and it’s always fascinating to learn the theories, implement them and witness their functionality in practical works (when they work).
Yukun Liu, class of 2022


Geert-Jan Huizing : témoignage filière TSIA

The Signal Processing for Artificial Intelligence course combines signal processing (sound analysis, signal compression, etc.) with data science (statistics, machine learning, etc.). Despite a strong mathematical emphasis (optimisation, statistics and a little time series), the numerous practical assignments make it a very applied subject: we worked on the data from the French National Great Debate, manipulated neural networks, or even did speech recognition. I never imagined that I would enjoy mathematics, Python, Numpy or Scipy so much during my preparatory studies!
Geert-Jan Huizing, class of 2020


Head: Roland Badeau
Head of international mobility: Giovanna Varni
Internship coordination: Marco Cagnazzo