Self-consistent learning of neural dynamical systems from noisy time series

• Main Authors: Wang, Zhe, Guet, Claude
• Other Authors: School of Physical and Mathematical Sciences
• Format: Article
• Language: English
• Published: 2022
• Subjects: Engineering::Computer science and engineering; Time Series Analysis; Dynamical Systems
• Online Access: https://hdl.handle.net/10356/162829
• Institution: Nanyang Technological University

We introduce a new method which, for a single noisy time series, provides unsupervised filtering, state space reconstruction, efficient learning of the unknown governing multivariate dynamical system, and deterministic forecasting. We construct both the underlying trajectories and a latent dynamical...