Speaker: Jean-Pierre Briot
Deep learning for music generation: Origins, successes and challenges
A growing area of application of the current wave of deep learning (the hyper-vitamined return of artificial neural networks) is the generation of creative content, notably the case of music (and also images and text). The motivation is in using machine learning techniques to automatically learn musical styles from arbitrary musical corpora and then to generate musical samples from the estimated distribution, with some degree of control over the generation. In this talk, we will at first analyze some early works from the late 1980s using artificial neural networks for music generation and how their pioneering contributions have prefigured recent techniques (e.g., hierarchical models and Deep Dream). We will then present some recent achievements using last generation architectures such as VAE, GAN and Transformer and analyze successes and challenges. Last, but not the least, will also address the issue of the authorship of the generated music and corresponding rights issues.
Jean-Pierre Briot is a senior researcher in computer science at CNRS (Centre National de la Recherche Scientifique) and Sorbonne Université in Paris, France. He is also permanent visiting professor at PUC-Rio and has recently been visiting professor at UNIRIO, both in Rio de Janeiro, Brazil. His general research interests are about the design of intelligent adaptive and cooperative software, at the crossing of artificial intelligence, distributed systems and software engineering, with various application fields such as Internet of Things, decision support systems and computer music. He is the first author of a recent reference book on the use of deep learning (artificial intelligence/machine learning) techniques for music generation. His current interest is focused on AI and music creativity.
Jean-Pierre Briot holds a masters in mathematics (1980), a doctorship (PhD) in computer science (1984) and an “habilitation à diriger des recherches” in computer science (1989), all from Université Pierre et Marie Curie (aka Paris VI, since 2018 renamed/merged as Sorbonne Université). He also holds degrees in music, music acoustics and Japanese language. He has been visiting researcher in various institutions (Kyoto University, Tokyo Institute of Technology, University of Illinois at Urbana-Champaign, University of Southern California, University of Tokyo…). He has advised more than 25 PhD students. He has edited 12 books or journal special issues. In 2010, he has created the CNRS permanent representation office in Rio de Janeiro, for scientific cooperation with Southern America.
For more details (including access to publications), please see http://www-desir.lip6.fr/~briot/cv/