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

We address the problem of safely solving complex bimanual robot manipulationtasks with sparse rewards. Such challenging tasks can be decomposed intosub-tasks that are accomplishable by different robots concurrently orsequentially for better efficiency. While previous reinforcement learningapproaches primarily focus on modeling the compositionality of sub-tasks, twofundamental issues are largely ignored particularly when learning cooperativestrategies for two robots: (i) domination, i.e., one robot may try to solve atask by itself and leaves the other idle; (ii) conflict, i.e., one robot caninterrupt another's workspace when executing different sub-taskssimultaneously, which leads to unsafe collisions. To tackle these two issues,we propose a novel technique called disentangled attention, which provides anintrinsic regularization for two robots to focus on separate sub-tasks andobjects. We evaluate our method on five bimanual manipulation tasks.Experimental results show that our proposed intrinsic regularizationsuccessfully avoids domination and reduces conflicts for the policies, whichleads to significantly more efficient and safer cooperative strategies than allthe baselines. Our project page with videos is athttps://mehooz.github.io/bimanual-attention.

DAIR: Disentangled Attention Intrinsic Regularization for Safe and Efficient Bimanual Manipulation