Machine Learning Makes Waves: A Better Way To Study Ocean Currents

A new machine-learning model developed by a multidisciplinary research team including computer scientists at MIT and oceanographers makes more accurate predictions about ocean currents than conventional models do. The model incorporates knowledge from fluid dynamics to better reflect the physics at work in ocean currents, and is more accurate at predicting currents and identifying divergences. This could help with tracking plastic pollution and oil spills, and aid in search and rescue.