Portfolio - Machine Learning
I am currently working on some ongoing projects. As this work is currently pending review for publication, only broad outlines will be provided below to maintain anonymity of the manuscript.
My main passion is in the modeling of patterns and trends, with long-term interests in two domains: Music and the arts, and Human-Computer Interaction, specifically in Adaptive Learning and Intelligent Tutoring Systems.
In the domain of Music Information Retrieval, these are my research directions:
I am most interested in understanding patterns in music and developing systems that use these patterns to understand and create music. Currently, I am investigating an alternative method of modeling information for Transformers for Automatic Music Generation of symbolic music. Most research in this domain focuses on certain classical instruments; I am interested in the generation of short passages of technical instrument playing that can be conditioned on style and difficulty. I am currently preparing a manuscript for conference submission; my next planned project will be looking at style transfer in symbolic music.
In future, I would also like to do further work on Music Information Retrieval, from a Data Science angle. With my musical background, I am particularly interested in the transcription of inharmonic multi-instrument polyphonic musical audio, which is an ongoing challenge in Automatic Music Transcription. Most research in this sector focuses on 'clean' instruments like the voice or piano. I am interested to investigate improved synthetic data generation or augmentation methods for pushing the limits of f0 estimation for distorted polyphonic audio.
Both of these have strong applications in producing customised Learning Experiences, in terms of content generation and assessment respectively.
Additionally, as part of my work with the Sound and Music Computing Lab, I am also currently working on pitch tracking and automated evaluation for pronunciation of tonal language for speech recognition, and conducting research into the applications of this in language learning and education. I am currently preparing a manuscript for conference submission for understanding the use of such educational technology for a specific demographic.