BCAM Scientific Seminar: Combining Data Assimilation and Neural ODEs for Learning Hybrid Models of Dynamics

The development of data-informed predictive models for dynamical systems is of widespread interest in many disciplines. Here, we present a unifying framework for blending mechanistic and machine-learning approaches for identifying dynamical systems from data. This framework is agnostic to the chosen machine learning model parameterization, and casts the problem in both continuous- and discrete-time. We will focus on recent developments that fuse data assimilation with auto-differentiable ODE solvers which, when combined, allow us to learn from noisy, partial observations. We will also present numerical experiments that demonstrate the efficacy of our approaches in Lorenz systems, as well as physiological systems. Joint work with Andrew Stuart.