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

Training long-horizon robotic policies in complex physical environments isessential for many applications, such as robotic manipulation. However,learning a policy that can generalize to unseen tasks is challenging. In thiswork, we propose to achieve one-shot task generalization by decoupling plangeneration and plan execution. Specifically, our method solves complexlong-horizon tasks in three steps: build a paired abstract environment bysimplifying geometry and physics, generate abstract trajectories, and solve theoriginal task by an abstract-to-executable trajectory translator. In theabstract environment, complex dynamics such as physical manipulation areremoved, making abstract trajectories easier to generate. However, thisintroduces a large domain gap between abstract trajectories and the actualexecuted trajectories as abstract trajectories lack low-level details and arenot aligned frame-to-frame with the executed trajectory. In a mannerreminiscent of language translation, our approach leverages a seq-to-seq modelto overcome the large domain gap between the abstract and executabletrajectories, enabling the low-level policy to follow the abstract trajectory.Experimental results on various unseen long-horizon tasks with different robotembodiments demonstrate the practicability of our methods to achieve one-shottask generalization.

Abstract-to-Executable Trajectory Translation for One-Shot Task Generalization