PHY Colloquium
Abstract: In this talk, I’ll share a personal perspective on where AI and machine learning can meaningfully support and accelerate scientific research, with concrete examples from our work in oceanography.
I’ll begin by highlighting opportunities to replace traditional methods with more efficient machine learning approaches, such as in data assimilation and surrogate modeling for ensemble-based space sampling. Next, I’ll discuss how AI can be used to refine model components, including updates to physical parameterizations like the K-Profile Parameterization for vertical mixing.
Finally, I’ll show how AI can help expand and enhance observational capabilities by leveraging numerical model outputs.