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

Robot manipulation requires a complex set of skills that need to be carefullycombined and coordinated to solve a task. Yet, most ReinforcementLearning (RL)approaches in robotics study tasks which actually consist only of a singlemanipulation skill, such as grasping an object or inserting a pre-graspedobject. As a result the skill ('how' to solve the task) but not the actual goalof a complete manipulation ('what' to solve) is specified. In contrast, westudy a complex manipulation goal that requires an agent to learn and combinediverse manipulation skills. We propose a challenging, highly under-actuatedpeg-in-hole task with a free, rotational asymmetrical peg, requiring a broadrange of manipulation skills. While correct peg (re-)orientation is arequirement for successful insertion, there is no reward associated with it.Hence an agent needs to understand this pre-condition and learn the skill tofulfil it. The final insertion reward is sparse, allowing freedom in thesolution and leading to complex emerging behaviour not envisioned during thetask design. We tackle the problem in a multi-task RL framework using ScheduledAuxiliary Control (SAC-X) combined with Regularized Hierarchical PolicyOptimization (RHPO) which successfully solves the task in simulation and fromscratch on a single robot where data is severely limited.

"What, not how": Solving an under-actuated insertion task from scratch