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

Recent research demonstrated that it is feasible to end-to-end trainmulti-task deep visuomotor policies for robotic manipulation using variationsof learning from demonstration (LfD) and reinforcement learning (RL). In thispaper, we extend the capabilities of end-to-end LfD architectures to objectmanipulation in clutter. We start by introducing a data augmentation procedurecalled Accept Synthetic Objects as Real (ASOR). Using ASOR we develop twonetwork architectures: implicit attention ASOR-IA and explicit attentionASOR-EA. Both architectures use the same training data (demonstrations inuncluttered environments) as previous approaches. Experimental results showthat ASOR-IA and ASOR-EA succeed ina significant fraction of trials incluttered environments where previous approaches never succeed. In addition, wefind that both ASOR-IA and ASOR-EA outperform previous approaches even inuncluttered environments, with ASOR-EA performing better even in cluttercompared to the previous best baseline in an uncluttered environment.

Accept Synthetic Objects as Real: End-to-End Training of Attentive Deep Visuomotor Policies for Manipulation in Clutter