Prof. Kate Helsen, Faculty of Music, Western University

Mar 3, 2023, 11:30am EC4-2101A

Until about a decade ago, it was generally believed that medieval musical manuscripts were unsuitable subjects for data scientists because they presented difficult problems that didn’t map intuitively onto use-cases in industry. Kate Helsen’s work with interdisciplinary and international projects has challenged this idea. Her Optical Neume Recognition Project (2011 – 2014) focused on document analysis and optimizing image binarization workflow for digital images of medieval sources. As a co-investigator in the SIMSSA project (Single Interface for Music Score Searching and Analysis) based at the Music Tech labs at McGill University (2015 – 2022), she worked from the musicological side to optimize searching and comparative analysis of manuscript digital images on the web with the online interface, ‘Cantus Ultimus’. For the past four years she was principal investigator on the Melodic Construction and Evolution research project at Western University using some 6,000 chant melodies, recently publishing a paper in the Empirical Musicology Review. In addition, she has supervised two capstone computer science undergraduate theses; the first involved a large melodic dataset, and the second, the creation of a transformer-based neural network to perform unsupervised classification of chant melodies by structural similarity. She is currently involved in the development of an xml schema specifically developed for medieval notation signs (called MEI: Music Encoding Initiative). She comes to you today with one question: What’s all this about Image A.I.?