Speaker: Ken Déguernel (PhD student)
Date: June 2, 2016
Abstract:
Automatic music improvisation systems based on the OMax paradigm use training over a one-dimensional sequence to generate original improvisation. First, we propose a system creating improvisation in a closer way to a human improviser where the intuition of a context is enriched with knowledge. This system combines a probabilistic model taking into account the multidimensional aspect of music trained on a corpus, with a factor oracle (structure used in OMax). The probabilistic model in constructed by interpolating sub-models and represents the knowledge of the system, while the factor oracle represents the context. The results show the potential of such a system to perform better navigation in the factor oracle, guided by the knowledge on several dimensions. Second, we propose another system using cluster graphs and message passing algorithms able to create communication between several factor oracle. Each factor oracle correspond to either a dimension or a virtual musician enabling the system to have a large range of application from multidimensional improvisation generation with a sense of anticipation to collective improvisation (free jazz, string quartet,…).
