Computer-Assisted Composition: Survey and hands-on with Calliope.
·
Workshop session title:
Computer-Assisted Composition: Survey and hands-on with Calliope.
Monday September 9th, 13:30-15:30.
List of contributors: Pr. Philippe Pasquier
Biography
Philippe Pasquier is a professor at Simon Fraser University's School of Interactive Arts and Technology, where he directs the Metacreation Lab for Creative AI. He leads a research-creation program around generative systems for creative tasks. As such, he is a scientist specialized in artificial intelligence since 1999, a software designer, a multidisciplinary media artist, an educator, and a community builder. His contributions bridge fundamental research on generative systems, machine learning, affective computing, and computer-assisted creativity, with applied research in the creative software industry, and artistic practice in media art, using generative systems.
https://philippepasquier.com/
Abstract
This 2h workshop will introduce, review and demo modern approaches and interfaces for AI-driven computer-assisted composition (CAC). In particular, the Calliope system (https://www.metacreation.net/projects/calliope/) will be detailed, and participants will be guided to generate musical content. This hands-on experience will be used to support a discussion on user experience, contrast the various workflow and design decisions, and envision the future of these systems.
Workshop Description
Computer-Assisted Composition (CAC) Workshop plan (2h):
Introduction to the Computer-Assisted Composition Workshop (10 mins)
Review of recent systems: with an emphasis on currently functional and accessible systems (40 mins)
Hands-on generation and listening session (30 mins).
Discussion and conclusion (20 mins).
Our recent survey covers 90+ CAC systems. With this tutorial, we want to synthesize this incredible diversity of approaches by demonstrating selected examples of systems representing current trends in currently available commercial and non-commercial CAC systems.
In a second part, we will introduce and demonstrate in more detail our free-to-use Calliope system built upon our multitrack music machines (mmm) models. The AIMC audience will be invited to try and generate content using Calliope, and this will be used to fuel a conversation on the pros and cons of these systems. Notions of bias, controllability, creative workflow, authorship and user experience will be introduced and discussed throughout.
Technical rider
Calliope is a browser-based interface that works on any computer, thus allowing every participant to test it out. The workshop will be given in person.
The ethics of machine learning can be questionable at times. The issues of big data and their copyright will be raised and examples of approaches relying on small data and/or model crafting will be highlighted.
Likewise, a lot of the field is culturally biased given the accessibility and quantity of western music in relation to other traditions.
Finally, some of the research presented involved human participants, and covered by SFU Office of Research Ethics approval.
The research presented and the presenter would like to acknowledge the support from the National Science and Engineering Research Council of Canada, the Social Science and Humanities Research Council, and Canada Council for the Arts.
Creative Commons Attribution 4.0 International License (CC-BY 4.0)
