Explore the context: optimal data collection for context-conditional dynamics models
- In this paper, we learn dynamics models for parametrized families of dynamical systems with varying properties. The dynamics models are formulated as stochastic processes conditioned on a latent context variable which is inferred from observed transitions of the respective system. The probabilistic formulation allows us to compute an action sequence which, for a limited number of environment interactions, optimally explores the given system within the parametrized family. This is achieved by steering the system through transitions being most informative for the context variable. We demonstrate the effectiveness of our method for exploration on a non-linear toy-problem and two well-known reinforcement learning environments.
Author: | Jan Achterhold, Jörg StücklerORCiDGND |
---|---|
Frontdoor URL | https://opus.bibliothek.uni-augsburg.de/opus4/107974 |
URL: | https://proceedings.mlr.press/v130/achterhold21a.html |
ISSN: | 2640-3498OPAC |
Parent Title (English): | Proceedings of The 24th International Conference on Artificial Intelligence and Statistics, 13-15 April 2021, Virtual |
Publisher: | ML Research Press |
Place of publication: | Maastricht |
Editor: | Arindam Banerjee, Kenji Fukumizu |
Type: | Conference Proceeding |
Language: | English |
Year of first Publication: | 2021 |
Release Date: | 2023/10/10 |
First Page: | 3529 |
Last Page: | 3537 |
Series: | Proceedings of Machine Learning Research (PMLR) ; 130 |
Institutes: | Fakultät für Angewandte Informatik |
Fakultät für Angewandte Informatik / Institut für Informatik | |
Fakultät für Angewandte Informatik / Institut für Informatik / Professur für Intelligente Perzeption in technischen Systemen |