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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Recent advances in deep reinforcement learning (RL) have demonstrated itspotential to learn complex robotic manipulation tasks. However, RL stillrequires the robot to collect a large amount of real-world experience. Toaddress this problem, recent works have proposed learning from expertdemonstrations (LfD), particularly via inverse reinforcement learning (IRL),given its ability to achieve robust performance with only a small number ofexpert demonstrations. Nevertheless, deploying IRL on real robots is stillchallenging due to the large number of robot experiences it requires. Thispaper aims to address this scalability challenge with a robust,sample-efficient, and general meta-IRL algorithm, SQUIRL, that performs a newbut related long-horizon task robustly given only a single video demonstration.First, this algorithm bootstraps the learning of a task encoder and atask-conditioned policy using behavioral cloning (BC). It then collectsreal-robot experiences and bypasses reward learning by directly recovering aQ-function from the combined robot and expert trajectories. Next, thisalgorithm uses the Q-function to re-evaluate all cumulative experiencescollected by the robot to improve the policy quickly. In the end, the policyperforms more robustly (90%+ success) than BC on new tasks while requiring notrial-and-errors at test time. Finally, our real-robot and simulatedexperiments demonstrate our algorithm's generality across different statespaces, action spaces, and vision-based manipulation tasks, e.g.,pick-pour-place and pick-carry-drop.

SQUIRL: Robust and Efficient Learning from Video Demonstration of Long-Horizon Robotic Manipulation Tasks