Artificial intelligence and music. Artificial Intelligence (AI) – specifically, deep learning models – is
developing rapidly and is increasingly being applied to creative tasks,
and not just music. The thrust of this research project has been about
the capacity of AI as a stimulator and assistant to human creativity.
- Submitting institution
-
Kingston University
- Unit of assessment
- 33 - Music, Drama, Dance, Performing Arts, Film and Screen Studies
- Output identifier
- 33-04-2076
- Type
- J - Composition
- Month
- February
- Year
- 2016
- URL
-
-
- Supplementary information
-
-
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- No
- Number of additional authors
-
-
- Research group(s)
-
-
- Proposed double-weighted
- Yes
- Double-weighted statement
- Presented here are the results of a 3-year multidisciplinary research project into an emerging research domain. The process started with applying machine-learning (ML) technology to a large data-set. Over the course of the project we created several ML models and developed new methods for evaluation of ML in relation to music, so that we could explore the capabilities of this cutting edge technology in creative practice and its wider implications for future developments in music.
- Reserve for an output with double weighting
- No
- Additional information
- Artificial Intelligence (AI) – specifically, deep learning models – is developing rapidly and is increasingly being applied to creative tasks, and not just music. The thrust of this research project has been about the capacity of AI as a stimulator and assistant to human creativity. Over a period of three years, the research addressed a range of questions – musical, technological, social and legal – through both creative and scholarly work. This multifaceted approach – integrating data science with music and grounding investigations in musical practices – is particularly important at this junction for a young and fast-moving domain.
Presented here are two ensemble compositions and three published articles. At the heart of this research enquiry is a close collaboration with engineering colleague Dr Bob Sturm. The two successful AHRC funding applications that supported and developed this research pathway were co-authored, as were the journal papers.
The material presented illustrates the range of methods used to evaluate the models: statistical benchmarking, music analysis, creative interrogation and expert solicitation. The range of research methods and outputs allowed the researchers to engage wider audiences, as evidenced through impact and engagement activities included. All three articles were conceived by Ben-Tal and Sturm and co-authored through continuous dialogue and collaboration. The two compositions (single-authored by Ben-Tal) used the models developed to generate composition material. They illustrate the different approaches that the creative side of this research enquiry can take, and their respective outcomes. The composition ‘Bastard Tune’ (Component #4) was premiered in London in May 2017, with further performances in London, Hamburg, DE, and Stockholm, SE. ‘Between the Lines’ (Component #5) was premiered in February 2018 at the University of Sussex, Falmer, UK, and performed again in London in October 2018.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -
