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

Machine Learning methods, such as those from the Reinforcement Learning (RL)literature, have increasingly been applied to robot control problems. However,such control methods, even when learning environment dynamics (e.g. as inModel-Based RL/control) often remain data-inefficient. Furthermore, thedecisions made by learned policies or the estimations made by learned dynamicmodels, unlike those made by their hand-designed counterparts, are not readilyinterpretable by a human user without the use of Explainable AI techniques.This has several disadvantages, such as increased difficulty both in debuggingand integration in safety-critical systems. On the other hand, in many roboticsystems, prior knowledge of environment kinematics and dynamics is at leastpartially available (e.g. from classical mechanics). Arguably, incorporatingsuch priors to the environment model or decision process can help address theaforementioned problems: it reduces problem complexity and the needs in termsof exploration, while also facilitating the expression of the decisions takenby the agent in terms of physically meaningful entities. Our aim with thispaper is to illustrate and support this point of view. We model a payloadmanipulation problem based on a real robotic system, and show that leveragingprior knowledge about the dynamics of the environment can lead to improvedexplainability and an increase in both safety and data-efficiency,leading tosatisfying generalization properties with less data.

Data-efficient, Explainable and Safe Payload Manipulation: An Illustration of the Advantages of Physical Priors in Model-Predictive Control