Grant: Discovery Grants Program - Individual

The research program in this Discovery grant proposal aims to develop innovative and efficient deep learning solutions for signal processing tasks required in hearing aid devices. Hearing loss is a severe problem affecting many Canadians and hearing aids are an efficient way of at least partially compensating for several types of hearing loss. Quality-of-life benefits for hearing aid owners have increased over the last decade, but recent surveys indicate that further developments are still required. Deep learning methods have recently shown good potential to provide significant improvements for some of the signal processing tasks required in hearing aids (e.g., sources detection, noise suppression, acoustic feedback canceling), but the complexity of the best performing methods is still too large for practical use in hearing aids. To achieve tangible progress in the proposed research program, this proposal includes both strategies to reduce the complexity of deep learning systems suitable for hearing aids signal processing tasks and directions to develop better-performing deep learning systems for the tasks. This Discovery grant will train HQP in the highly in-demand field of machine learning for signal processing, with clear economic and healthcare benefits to Canada.