Predictability Enables Parallelization of Nonlinear State Space Models

The rise of parallel computing hardware has made it increasingly important to understand which nonlinear state space models can be efficiently parallelized.