Collapsed Amortized Variational Inference for Switching Nonlinear Dynamical Systems

10/21/2019

∙

by   Zhe Dong, et al.

∙

0

∙

share

We propose an efficient inference method for switching nonlinear dynamical systems. The key idea is to learn an inference network which can be used as a proposal distribution for the continuous latent variables, while performing exact marginalization of the discrete latent variables. This allows us to use the reparameterization trick, and apply end-to-end training with stochastic gradient descent. We show that the proposed method can successfully segment time series data (including videos) into meaningful "regimes", by using the piece-wise nonlinear dynamics.