Deezer Research actively collaborates with international academic partners through dedicated research
initiatives, data sharing and student training.
We also participate and lead open-source code initiatives of research tools.
Le projet GEM vise à décrire les différences de représentation et de traitement existant entre les femmes et les hommes dans les médias, en se fondant sur l’analyse automatique de gros volumes de données en langue française contenus dans les collections de l’INA et de Deezer : TV, radio, presse écrite et collections musicales.
Started in June 2016, this French National Research Agency (ANR) project focuses on applying
next generation machine learning methods to spatio-temporal series. The consortium is
composed of two academics, UPMC-LIP6-Paris and
UJF-LIG-Grenoble, and one industrial,
RECORDS is a collaborative research project funded through an ANR grant (2020-23). The goal is to improve our understanding of (i) the diversity of users practices and consumptions on streaming platforms (ii) the effects of manual and algorithmic content recommendation (iii) the space-time diffusion of music.
Spleeter is the Deezer source separation library
with pretrained models written in Python and using
Tensorflow. It makes it easy to train music source
separation models (assuming you have a dataset of isolated sources), and provides already
trained state of the art models for performing various flavours of separation.
The WASABI ANR project, started in January 2017, aims at the construction of a 2 million song knowledge
base that combines metadata collected from music databases on the Web, metadata resulting from the analysis of
song lyrics, and metadata resulting from the audio analysis.