Bandeau image

The « STARS » project

The « STARS » project

  • The project
  • The partners
Large-scale Computational Analysis of Audio Music : STatistical relational ARtificial intelligence for Structure discovery
The project
Corps de texte

The development of computer hardware technology and the proliferation of online music collections have sustained the development of Artificial Intelligence techniques for music research in several directions. The objective of this project is to contribute to the development of new strategies and technologies for processing digital music and enabling interaction/access to online music collections, and specifically for computational music structure analysis. Traditional approaches for music processing are not able to cope with its rich, highly structured, complex relational structure, and its inherent uncertainty and heterogeneity. We propose to address several limitations of current approaches by taking advantage of recent developments in the area of Statistical Relational Artificial Intelligence that integrate probability, logic and deep learning. This project is strongly interdisciplinary and will bring together efforts from various scientific domains (signal processing, computer science, statistical relational learning, deep learning, high dimensional data analysis), but also knowledge from humanistic disciplines such as music analysis, musicology and jazz music studies. It will benefit from the combination of the diverse fields of expertise from the different members involved in the team. Beside the development of innovative models for music processing it will address general AI methodological questions, and it will also contribute to the interaction with related digital humanities disciplines.

The partners
Corps de texte

Project leaders

H.-C. Crayencour (L2S - CNRS)

Expert in statistical relational learning and music information retrieval.


S. Essid (LTCI, Télécom Paris, Institut Polytechnique de Paris)

Expert in structured prediction, representation learning, and audio signal processing.


M. Kowalski (L2S - Université Paris-Saclay)

Expert in structured sparse approximations and data decomposition methods.



GPI team, at Laboratoire des Signaux et Systèmes (L2S) - CentraleSupélec

3, Rue Joliot-Curie

91192 Gif-sur-Yvette


S²A team, Laboratoire Traitement et Communication de l’Information (LTCI) - Télécom Paris

19 place Marguerite Perey

F-91120 Palaiseau