English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

This paper deals with robotic lever control using Explainable DeepReinforcement Learning. First, we train a policy by using the DeepDeterministic Policy Gradient algorithm and the Hindsight Experience Replaytechnique, where the goal is to control a robotic manipulator to manipulate alever. This enables us both to use continuous states and actions and to learnwith sparse rewards. Being able to learn from sparse rewards is especiallydesirable for Deep Reinforcement Learning because designing a reward functionfor complex tasks such as this is challenging. We first train in the PyBulletsimulator, which accelerates the training procedure, but is not accurate onthis task compared to the real-world environment. After completing the trainingin PyBullet, we further train in the Gazebo simulator, which runs more slowlythan PyBullet, but is more accurate on this task. We then transfer the policyto the real-world environment, where it achieves comparable performance to thesimulated environments for most episodes. To explain the decisions of thepolicy we use the SHAP method to create an explanation model based on theepisodes done in the real-world environment. This gives us some results thatagree with intuition, and some that do not. We also question whether theindependence assumption made when approximating the SHAP values influences theaccuracy of these values for a system such as this, where there are somecorrelations between the states.

Robotic Lever Manipulation using Hindsight Experience Replay and Shapley Additive Explanations