In this presentation, I will be discussing the main findings from my PhD project on speech enhancement
 for hearing aids. In noisy environments, speech enhancement is crucial for improving speech intelligibility and quality for hearing aid (HA) users. However, current multi-microphone noise reduction
 techniques often rely on assumptions about the acoustic environment and
 the desired speaker (i.e., the target) that may not hold in real-world scenarios.

For example, beamformers are often configured to enhance sound from the frontal direction with respect to the HA user, assuming that the target is positioned in front of the user. This may provide suboptimal performance in some acoustic environments. In addition, highly non-stationary and modulating noise sources such as competing speakers can make it challenging for current speech enhancement solutions to distinguish between the target speech and competing speech signals since
 they share similar temporal and spatial characteristics.

Furthermore, current speech enhancement algorithms lack an understanding of the user's intention, making them less effective in situations where users are engaged in conversations or interacting with the environment. To address these limitations, my PhD project proposes a user-symbiotic approach to speech enhancement for hearing aids. The user-symbiotic approach leverages the user's turn-taking behavior and own voice to provide valuable information about what is desired and what is not for noise reduction algorithms. By exploiting this information, we can develop noise reduction algorithms that are more effective at improving speech intelligibility and quality over current uninformed algorithms.

In this presentation, I will provide an overview of the proposed approach, including the underlying algorithms and experimental results. I will also discuss the potential impact of the approach and its future directions, highlighting the need for further research and development in this critical area of hearing aid technology.