Switching Dynamical Systems with Deep Neural Networks

Author: C. Ojeda, B. Georgiev, K. Cvejoski, J. Schuecker, C. Bauckhage, R. Sanchez
Journal: 2020 25th International Conference on Pattern Recognition (ICPR)
Year: 2020

Citation information

C. Ojeda, B. Georgiev, K. Cvejoski, J. Schuecker, C. Bauckhage, R. Sanchez,
2020 25th International Conference on Pattern Recognition (ICPR),
2020,
6305-6312,
IEEE,
Milan, Italy,
https://doi.org/10.1109/ICPR48806.2021.9412566

The problem of uncovering different dynamical regimes is of pivotal importance in time series analysis. Switching dynamical systems provide a solution for modeling physical phenomena whose time series data exhibit different dynamical modes. In this work we propose a novel variational RNN model for switching dynamics allowing for both non-Markovian and non-linear dynamical behavior between and within dynamic modes. Attention mechanisms are provided to inform the switching distribution.