Switching Linear Dynamics for Variational Bayes Filtering
[edit]
Proceedings of the 36th International Conference on Machine Learning, PMLR 97:553562, 2019.
Abstract
System identification of complex and nonlinear systems is a central problem for model predictive control and modelbased reinforcement learning. Despite their complexity, such systems can often be approximated well by a set of linear dynamical systems if broken into appropriate subsequences. This mechanism not only helps us find good approximations of dynamics, but also gives us deeper insight into the underlying system. Leveraging Bayesian inference, Variational Autoencoders and Concrete relaxations, we show how to learn a richer and more meaningful state space, e.g. encoding joint constraints and collisions with walls in a maze, from partial and highdimensional observations. This representation translates into a gain of accuracy of learned dynamics showcased on various simulated tasks.
Related Material


