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

The ability to leverage shared behaviors between tasks is critical forsample-efficient multi-task reinforcement learning (MTRL). While prior methodshave primarily explored parameter and data sharing, direct behavior-sharing hasbeen limited to task families requiring similar behaviors. Our goal is toextend the efficacy of behavior-sharing to more general task families thatcould require a mix of shareable and conflicting behaviors. Our key insight isan agent's behavior across tasks can be used for mutually beneficialexploration. To this end, we propose a simple MTRL framework for identifyingshareable behaviors over tasks and incorporating them to guide exploration. Weempirically demonstrate how behavior sharing improves sample efficiency andfinal performance on manipulation and navigation MTRL tasks and is evencomplementary to parameter sharing. Result videos are available athttps://sites.google.com/view/qmp-mtrl.

Efficient Multi-Task Reinforcement Learning via Selective Behavior Sharing