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 field of cooperative multi-agent reinforcement learning (MARL) has seenwidespread use in addressing complex coordination tasks. While valuedecomposition methods in MARL have been popular, they have limitations insolving tasks with non-monotonic returns, restricting their generalapplication. Our work highlights the significance of joint intentions incooperation, which can overcome non-monotonic problems and increase theinterpretability of the learning process. To this end, we present a novel MARLmethod that leverages learnable joint intentions. Our method employs ahierarchical framework consisting of a joint intention policy and a behaviorpolicy to formulate the optimal cooperative policy. The joint intentions areautonomously learned in a latent space through unsupervised learning and enablethe method adaptable to different agent configurations. Our results demonstratesignificant performance improvements in both the StarCraft micromanagementbenchmark and challenging MAgent domains, showcasing the effectiveness of ourmethod in learning meaningful joint intentions.

Multi-Agent Cooperation via Unsupervised Learning of Joint Intentions