Project Title: UPF-BMAT Chair on Artificial Intelligence and Music
Creator: Xavier Serra
Affiliation: Universitat Pompeu Fabra
Funder: Secretaría de Estado de Digitalización e Inteligencia Artificial, Ministerio de
Transformación Digital y Función Pública, and European Union-Next Generation EU
Grant number: TSI-100929-2023-1
URL: https://www.upf.edu/web/mtg/catedra-ia-musica
 

Project abstract:

The UPF-BMAT Chair on Artificial Intelligence and Music is an initiative of the Music Technology Group (MTG) at the Department of Engineering of Universitat Pompeu Fabra (UPF), in collaboration with BMAT, a company specializing in music technology. The Chair is part of the National Artificial Intelligence Strategy (ENIA), co-funded by the Ministry of Economic Affairs and Digital Transformation of Spain and BMAT Music Innovators.

The director of the Chair is Prof. Xavier Serra, director of the MTG, and it involves researchers and staff from both the MTG and BMAT, establishing collaborations with institutions and experts in the music sector throughout Spain.

The Chair has an initial budget of 1.2 million euros for activities during the first 4 years. The primary objective of the Chair is to assist the music sector in harnessing the potential of AI by developing practical applications and training professionals capable of leading the
transformation and renewal of the music industry in Spain.

The Chair will be a valuable tool for identifying and capitalizing on opportunities presented by AI technologies in music creation, production, distribution, protection, licensing and education. Furthermore, it will help educate society, music consumers and music industry
stakeholders about the possibilities and limitations of AI so that they can effectively use and benefit from these technologies.

The Chair will undertake the following activities within the music sector:

1. Doctoral, master's, and continuous education in AI and music. It will train future professionals and assist active professionals in updating their knowledge in AI-related topics.

2. Fundamental research in AI methodologies for music creation, production, distribution, protection, licensing and education. The Chair will conduct research with academic and industrial impact, collaborating with sector stakeholders to address their needs. Emphasizing an open science culture, research will focus on interpretability, transparency, diversity, and accessibility of developed AI methodologies and systems.

3. Applied research, developing software, datasets, and prototypes beneficial to the sector. The Chair will create advanced and industrially scalable technologies, promoting open innovation by facilitating public-private collaboration and collaboration among different sector agents.

4. Dissemination of AI and music research, both conducted by the Chair and internationally. It will create open content available on the Chair's website and promote AI and music topics through social media. The Chair will maintain an open dialogue with the sector and the public on ethical issues arising from the use of AI methodologies in music.

5. Technological transfer of obtained results. The Chair will help the Spanish music sector incorporate AI technologies into its processes, promoting the transfer of AI technologies through contracts with companies, technology licenses, technology consulting, and the creation of spin-offs and start-ups.

The objectives of the Chair align with those of the National Artificial Intelligence Strategy (ENIA).
The planned activities fall within the first 4 strategic axes defined in ENIA: (1) Promoting scientific research, technological development, and innovation in AI, (2) Fostering the development of digital capabilities, enhancing national talent, and attracting global talent, (3)
Developing data platforms and technological infrastructures supporting AI, and (4) Integrating AI into value chains to transform the economic fabric.

Start date: 01-04-2023
End date: 31-03-2027

1. Data summary

The Chair on AI and Music will generate and reuse data as part of its research, training, and technology transfer activities focused on the development and evaluation of artificial intelligence technologies applied to music.

The types of data generated and used include:

  • Audio recordings: musical performances by musicians or students.

  • Annotations: time stamps, semantic tags, and expert or crowdsourced analyses.

  • Listening test results: questionnaires, user judgments, and interaction logs.

  • Symbolic music data: digitized scores (MusicXML, MIDI).

  • Derived data: trained AI models, evaluation metrics, and usage logs.

These data will support technology development in three main areas:

  • Music creation and production.

  • Music distribution and monitoring.

  • Music education supported by AI.

2. Applicable Standards and Ethical Principles

The collection and processing of personal data complies with the EU General Data Protection Regulation (GDPR, Regulation 2016/679) and is supervised by the UPF's Institutional Research Ethics Committee (CIREP-UPF).

No sensitive categories of data will be collected. All personal data are anonymized prior to analysis and publication. Informed consent is mandatory and documented. Data are stored on secure servers at UPF with restricted access.

3. Data Documentation and Quality

The data generated will include:

  • Standardized metadata (Dublin Core, schema.org, or similar).

  • Format specifications (WAV, MP3, CSV, JSON, XML, MusicXML, MIDI).

  • Annotation protocols.

  • Provenance and versioning logs for datasets and models.

Version control systems (e.g., Git) and peer-review procedures will ensure data and annotation quality.

4. Storage and Backup During the Project

Data will be stored and backed up using the IT infrastructure of Universitat Pompeu Fabra:

  • Secure internal servers (with access control).

  • Regular automated backups.

  • Working repositories organized by data type and working group.

Data integrity, protection from unauthorized access, and secure storage will be guaranteed following institutional protocols.

5. Long-term Preservation and Reuse

Data will be preserved for at least 5 years after project completion, in line with UPF policy and the State Plan's open-access requirements.

Datasets without personal data or with fully anonymized content will be published under open licenses (Creative Commons BY or similar) on repositories such as:

  • Zenodo (MTG-UPF community)

  • Freesound.org

  • UPF Digital Repository

6. Data Sharing and Open Access

The Chair will promote data sharing aligned with the FAIR principles:

  • Findable: persistent identifiers and indexed metadata.

  • Accessible: open access via HTTPS with clear licensing.

  • Interoperable: use of standard formats (WAV, MIDI, JSON, MusicXML).

  • Reusable: full documentation, citability, and traceability.

Research results and related data will also be shared through open-access publications and via the Chair’s website and academic networks.

7. Responsibilities

  • The Principal Investigator, Xavier Serra, is responsible for implementing the data management plan.

  • Each sub-project (training, research, transfer) will designate a data coordinator.

  • The UPF Data Protection Officer will ensure legal compliance.

8. Resources

The MTG-UPF team has extensive experience in managing musical and scientific data. Existing technical infrastructure (servers, repositories, version control tools) will be used, and part of the budget will be allocated to:

  • Technical and ethical support for data management.

  • Development of documentation and anonymization tools.

  • Open-access publication costs.