Computer music designer, musician, and researcher Jérôme Nika (FR) will present the generative agents / software instruments DYCI2 that he develops in collaboration with Ircam, and in interaction with expert improvisers. These agents offer a continuum of strategies going from pure autonomy to meta-composition thanks to an abstract “scenario” structure. More than 60 artistic performances have brought these tools into play since 2016, particularly in improvised music (Steve Lehman, Bernard Lubat, Benoît Delbecq, Rémi Fox) and contemporary music (Pascal Dusapin, Marta Gentilucci). Daniele Ghisi will explore, how machines learn to produce music. A concept-album of generative music, a collection of musical monsters exploring sample-by-sample deep learning generative models, inspired by Mary Shelley’s Frankenstein and used in Jean François Peyret’s eponymous theatrical piece.