Mixed Initiative Co-Creative Design for Long-Term Human-AI Musical Partnerships – MICCDroP

Current Human-Artificial intelligence (Ai) co-creative systems, also known as Mixed Initiative Co-Creative systems (MICC) have opened new avenues in music generation. These systems facilitate novel creative processes and modes of interactions. However, these systems lack the ability for long-term adaptation between user and machine. Current ways of adapting are either unilateral, where the user adaptively learns to operate the Ai system, or self-engineered, where musician-engineers modify their own systems over time. The 3-years MICCDroP project aims to address this limitation by developing Ai methods designed for lifelong learning, and evaluating the impact of such methods through musicology and creativity theory lenses. This places our project squarely within the field of digital humanities, aligning with the theme digital approaches for artistic creation, computer assisted music, and personalisation and adaptation.
The project is organised around 4 work-packages:
WP1: Lifelong Learning for MICC music generation: aiming at applying and developing lifelong learning techniques for machine learning music generation systems with a focus on agent personalisation.
WP2: Exploring the creative aspects of lifelong MICC: aiming at critically evaluate the impact of the lifelong learning through the lens of Computational Creativity and Ai ethics.
WP3: Artistic experimentation: aiming at carrying collaborations with expert musicians through experimental sessions with interviews and ethnographical studies.
WP4: Management and dissemination.
This project will be the first application of lifelong learning methods to music, and aim at expending our knowledge on co-creativity; music being a great field for this as it is a complex human activity with deep cultural and cognitive impact, and high interactivity.