Yu, Chen and Rosendo, Andre (2021) Risk-Aware Model-Based Control. Frontiers in Robotics and AI, 8. ISSN 2296-9144
![[thumbnail of pubmed-zip/versions/1/package-entries/frobt-08-617839.pdf]](http://publish7promo.com/style/images/fileicons/text.png)
pubmed-zip/versions/1/package-entries/frobt-08-617839.pdf - Published Version
Download (4MB)
Abstract
Model-Based Reinforcement Learning (MBRL) algorithms have been shown to have an advantage on data-efficiency, but often overshadowed by state-of-the-art model-free methods in performance, especially when facing high-dimensional and complex problems. In this work, a novel MBRL method is proposed, called Risk-Aware Model-Based Control (RAMCO). It combines uncertainty-aware deep dynamics models and the risk assessment technique Conditional Value at Risk (CVaR). This mechanism is appropriate for real-world application since it takes epistemic risk into consideration. In addition, we use a model-free solver to produce warm-up training data, and this setting improves the performance in low-dimensional environments and covers the shortage of MBRL’s nature in the high-dimensional scenarios. In comparison with other state-of-the-art reinforcement learning algorithms, we show that it produces superior results on a walking robot model. We also evaluate the method with an Eidos environment, which is a novel experimental method with multi-dimensional randomly initialized deep neural networks to measure the performance of any reinforcement learning algorithm, and the advantages of RAMCO are highlighted.
Item Type: | Article |
---|---|
Subjects: | Euro Archives > Mathematical Science |
Depositing User: | Managing Editor |
Date Deposited: | 30 Jun 2023 04:06 |
Last Modified: | 03 Oct 2023 09:12 |
URI: | http://publish7promo.com/id/eprint/2857 |