Blogpost by Gissel Velarde, researcher in Music Information Retrieval and WiMIR mentor, summarising her PhD work.
In April this year, I defended my thesis entitled: Convolutional Methods for Music Analysis, available here.
This work introduces convolution, its relevance for perceptual tasks, and its effect on music analysis in applications to music segmentation, pattern discovery and classification. The methodology we have followed was to systematically study and evaluate the effect of convolution (filtering) and other processing techniques together with machine learning algorithms, from k-nearest neighbours, single linkage, support vector machines to convolutional neural networks.
The novel convolution-based methods for music analysis presented in my thesis have been developed together with my supervisors Associate Professor David Meredith, Aalborg University and Senior Lecturer Tillman Weyde City, University of London, as well as in collaboration with researchers from The Austrian Research Institute for Artificial Intelligence: Carlos Cancino Chacón and Maarten Grachten.
Picture of PhD Defense of Gissel Velarde, April 2017, Aalborg University
Gissel Velarde completed her PhD studies in computer science at Aalborg University, supported by a scholarship from the Department of Architecture Design and Media Technology, Aalborg University, and partially supported by the European Commission, FET grant number 610859. She also holds a Masters degree in Electronic Systems and Engineering Management from the Südwestfallen University of Applied Sciences, supported by a DAAD scholarship. Her Licenciatura degree in Systems Engineering was obtained from the Universidad Católica Boliviana. She was a research member of the European Commission Project “Learning to Create” (Lrn2Cre8).
Before dedicating to technology, Velarde studied piano at the Conservatorio Plurinacional de Música in La Paz, Bolivia and won as a pianist, several prices and honors.
During her doctoral studies at Aalborg University, she published research papers on computational methods for music analysis. She was teaching assistant on the Master of Science program in Sound and Music Computing and supervised various projects of the Bachelor program in Medialogy.