Classical music is one of the greatest treasures of Europe’s cultural heritage. Although a historical genre, it is continually (re)interpreted and revitalised through musical performance. Today, most of the classical repertoire is in the public domain; massive numbers of scores and recordings are now available in online community-contributed repositories actively used by scholars and musicians. Technology offers ways to enrich and contextualise this repertoire, so that users might better understand and appreciate it. However, due to varying data quality and scale, this does not happen automatically for public-domain resources. Amidst a deluge of data, relevant associations across repositories and modalities (e.g. from scores to recordings) still have to be made manually, while insights by previous users are not explicitly stored for future users to learn from. It is thus impossible to get comprehensive insight into the full wealth of our musical cultural heritage.
TROMPA aims at changing this by massively enriching and democratising our publicly available musical heritage through a user-centred co-creation setup. For analysing and linking music data at scale, the project studies how to improve state-of-the-art music analysis technology with the help of music-loving citizens (including the large scene of amateur performers) that annotate the data according to their personal expertise, and provide feedback on algorithmic results, and annotating the data according to their personal expertise.
Specific Topics within TROMPA
Within TROMPA, the two PhD candidates will study, design, and develop:
1) novel computational methods that exploit the availability and expertise (at varying levels) of the crowd to unlock knowledge and express own perspectives on music material; and
2) novel framework, exploiting the added value of human annotations, for the continuous evaluation of multimodal music analysis algorithms.
Delft University of Technology
We are looking for two candidates who can meet the following requirements:
– has an MSc in Computer Science, Statistics, or a closely related field;
– followed MSc courses and has research experience (e.g. at Master thesis level) in the at least some of the following areas: human computation, crowdsourcing, (multimedia) information retrieval, machine learning, user modelling, data mining, human-computer interaction.
– programming skills in modern languages;
– is a team player;
– enjoys research and is self-motivated;
– has good communication skills in English.
About TU Delft
Delft University of Technology in the Netherlands (TU Delft) is a modern university with a rich tradition. Its eight faculties and over 30 English-language Master programmes are at the forefront of technological development, contributing to scientific advancement in the interests of society.
Ranked among the top universities of technology in Europe (#18, QS 2010) TU Delft’s excellent research and education standards are backed by outstanding facilities, research institutes and research schools. TU Delft maintains close links with (inter)national industry, a strategic alliance contributing to the relevance of its academic programmes and career prospects for its graduates.
Society is our continuous incentive for research. We carry out research to find solutions for society’s present and future demands. Fundamental research is part of this, because we aim to find solutions for tomorrow’s problems. Health, energy, environment and infrastructures & mobility are today's major social issues. That's why TU De