The project and state of MIR

Music Information Processing (MIP), also known as Music Information Research (MIR), involves the use of information processing methodologies to understand and model music and to develop products and services for creation, distribution, and interaction with music and music-related information. MIR has reached a state of maturity where there are standard methods for most music information processing tasks, but as these have been developed and tested on small datasets, they tend to suffer from three main weaknesses:

  • not being robust to musical differences
  • not generalizing across different use contexts
  • not being scalable to industrial scale datasets.

Moreover, the interdisciplinary nature of MIR presents a challenge for researchers entering the field, who need to develop expertise in mathematical, algorithmic, musical and perceptual areas, plus generic skills such as software development, data management, innovation, collaboration, understanding industry needs, and reproducible research.

To meet this need, and to train a new generation of researchers who are aware of, and can tackle, these challenges, we bring together leading MIR groups and a wide range of industrial and cultural stakeholders to create a multidisciplinary, transnational and cross-sectoral European Training Network for MIR researchers, in order to contribute to Europe’s leading role in this field of scientific innovation, and accelerate the impact of innovation on European products and industry.

The scientific focus of MIP-Frontiers is motivated by our conviction that further substantial advances in MIR will require approaches of a new quality: MIR algorithms need to become more aware of musically and perceptually relevant concepts, they need to be based on solid musicological principles, and they need to take into account and exploit a multitude of additional information sources - including not only large repositories of audio material for unsupervised feature learning, but also multimodal datasets containing audio, video, scores, and textual and social user data from various sources. User needs, additional application knowledge, and real-world constraints will be brought in by the non-academic partners and stakeholders, of which this network has a broad and diverse array.

MIP-Frontiers network comprises 4 academic and 3 non-academic Beneficiaries, plus 9 Partners, and will train 15 researchers in a range of university-based and industry-based PhDs, all of which involve cross-sectoral training via secondments and network-wide events. The researchers will develop breadth in the fields that make up MIR and in transferable skills, whilst gaining deep knowledge and skills in their own area of specialty. They will learn to perform collaborative research, to think entrepreneurially and to exploit their research in new ways that benefit the European industry and society.

MIP-Frontiers explicitly brings together bottom-up (data-driven) information processing with top-down (knowledge-driven) processing, moderating both of these in a context of the usage (user-driven) in order to advance the state of the art in MIR. Furthermore, an Open Science approach will be adopted, and, wherever possible, the training resources, procedures, and reports will be released in order to help future training networks to learn the most from our experiences, and contribute to an improvement in the training of researchers in a wider range of fields.