Swimming with deep learning

The study of micro-organisms' propulsion has intrinsic relevance for the development of micro-robots designed for targeted drug delivery and as a foundation for further studies on hydrodynamic interactions between micro-organisms in complex environments. Numerical simulations have been used extensively to investigate micro-organisms' locomotion. Recently, physics-informed neural networks (PINNs) have shown promise for approximating solutions to differential equations that govern various physical problems. In this talk, we evaluate the effectiveness of using PINNs to predict the low Reynolds dynamics that characterize the propulsion of micro-organisms.