A flexible state space model for learning nonlinear dynamical systems
This work addresses the challenge of modeling complex nonlinear dynamics for researchers and practitioners in fields like control systems or time-series analysis, but it appears incremental as it builds on existing state-space and Gaussian process methods.
The authors tackled the problem of learning nonlinear dynamical systems by proposing a flexible state space model with basis function expansions and Gaussian process priors, achieving promising results on a classical benchmark and real data.
We consider a nonlinear state-space model with the state transition and observation functions expressed as basis function expansions. The coefficients in the basis function expansions are learned from data. Using a connection to Gaussian processes we also develop priors on the coefficients, for tuning the model flexibility and to prevent overfitting to data, akin to a Gaussian process state-space model. The priors can alternatively be seen as a regularization, and helps the model in generalizing the data without sacrificing the richness offered by the basis function expansion. To learn the coefficients and other unknown parameters efficiently, we tailor an algorithm using state-of-the-art sequential Monte Carlo methods, which comes with theoretical guarantees on the learning. Our approach indicates promising results when evaluated on a classical benchmark as well as real data.