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

Fleets of robots ingest massive amounts of streaming data generated byinteracting with their environments, far more than those that can be stored ortransmitted with ease. At the same time, we hope that teams of robots canco-acquire diverse skills through their experiences in varied settings. How canwe enable such fleet-level learning without having to transmit or centralizefleet-scale data? In this paper, we investigate distributed learning ofpolicies as a potential solution. To efficiently merge policies in thedistributed setting, we propose fleet-merge, an instantiation of distributedlearning that accounts for the symmetries that can arise in learning policiesthat are parameterized by recurrent neural networks. We show that fleet-mergeconsolidates the behavior of policies trained on 50 tasks in the Meta-Worldenvironment, with the merged policy achieving good performance on nearly alltraining tasks at test time. Moreover, we introduce a novel robotic tool-usebenchmark, fleet-tools, for fleet policy learning in compositional andcontact-rich robot manipulation tasks, which might be of broader interest, andvalidate the efficacy of fleet-merge on the benchmark.

Fleet Policy Learning via Weight Merging and An Application to Robotic Tool-Use