This project involves augmenting the existing physics with state-of-the-art generative AI technologies, including those based on score-matching and diffusion, to produce much more powerful models for acoustic waveforms and propagation in underwater settings. It will also investigate representative applications, to demonstrate the value of these models and the associated signal processing algorithms.

Through this SuperUROP project, I hope to develop my skills as a researcher and gain a deeper understanding of how we can leverage generative AI to solve real-world problems. My extensive course work in machine learning, as well as past UROP and teaching experience, has prepared me to creatively adapt and build upon existing diffusion model paradigms for this novel application.