Timezone: »
System identification of complex and nonlinear systems is a central problem for model predictive control and model-based 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 high-dimensional observations. This representation translates into a gain of accuracy of learned dynamics showcased on various simulated tasks.
Author Information
Philip Becker-Ehmck (Volkswagen Group)
Jan Peters (TU Darmstadt)
Patrick van der Smagt (Volkswagen Group)
Related Events (a corresponding poster, oral, or spotlight)
-
2019 Oral: Switching Linear Dynamics for Variational Bayes Filtering »
Tue. Jun 11th 10:00 -- 10:05 PM Room Room 201
More from the Same Authors
-
2021 : Exploration via Empowerment Gain: Combining Novelty, Surprise and Learning Progress »
Philip Becker-Ehmck · Maximilian Karl · Jan Peters · Patrick van der Smagt -
2022 : Probabilistic Dalek - Emulator framework with probabilistic prediction for supernova tomography »
Wolfgang Kerzendorf · Nutan Chen · Patrick van der Smagt -
2022 : Local distance preserving autoencoders using continuous kNN graphs »
Nutan Chen · Patrick van der Smagt · Botond Cseke -
2020 Poster: Learning Flat Latent Manifolds with VAEs »
Nutan Chen · Alexej Klushyn · Francesco Ferroni · Justin Bayer · Patrick van der Smagt