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

End-to-end learning robotic manipulation with high data efficiency is one ofthe key challenges in robotics. The latest methods that utilize humandemonstration data and unsupervised representation learning has proven to be apromising direction to improve RL learning efficiency. The use of demonstrationdata also allows "warming-up" the RL policies using offline data with imitationlearning or the recently emerged offline reinforcement learning algorithms.However, existing works often treat offline policy learning and onlineexploration as two separate processes, which are often accompanied by severeperformance drop during the offline-to-online transition. Furthermore, manyrobotic manipulation tasks involve complex sub-task structures, which are verychallenging to be solved in RL with sparse reward. In this work, we propose aunified offline-to-online RL framework that resolves the transition performancedrop issue. Additionally, we introduce goal-aware state information to the RLagent, which can greatly reduce task complexity and accelerate policy learning.Combined with an advanced unsupervised representation learning module, ourframework achieves great training efficiency and performance compared with thestate-of-the-art methods in multiple robotic manipulation tasks.

Efficient Robotic Manipulation Through Offline-to-Online Reinforcement Learning and Goal-Aware State Information