Title: Rodent: Relevance determination in ODE

Abstract: From a set of observed trajectories of a partially observed system, we aim to learn its underlying (physical) process without having to make too many assumptions about the generating model. We start with a very general, over-parameterized ordinary differential equation (ODE) of order N and learn the minimal complexity of the model, by which we mean both the order of the ODE as well as the minimum number of non-zero parameters that are needed to solve the problem. The minimal complexity is found by combining the Variational Auto-Encoder (VAE) with Automatic Relevance Determination (ARD) to the problem of learning the parameters of an ODE which we call Rodent. We show that it is possible to learn not only one specific model for a single process, but a manifold of models representing harmonic signals in general.