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

Learning communication strategies in cooperative multi-agent reinforcementlearning (MARL) has recently attracted intensive attention. Early studiestypically assumed a fully-connected communication topology among agents, whichinduces high communication costs and may not be feasible. Some recent workshave developed adaptive communication strategies to reduce communicationoverhead, but these methods cannot effectively obtain valuable information fromagents that are beyond the communication range. In this paper, we consider arealistic communication model where each agent has a limited communicationrange, and the communication topology dynamically changes. To facilitateeffective agent communication, we propose a novel communication protocol calledAdaptively Controlled Two-Hop Communication (AC2C). After an initial localcommunication round, AC2C employs an adaptive two-hop communication strategy toenable long-range information exchange among agents to boost performance, whichis implemented by a communication controller. This controller determineswhether each agent should ask for two-hop messages and thus helps to reduce thecommunication overhead during distributed execution. We evaluate AC2C on threecooperative multi-agent tasks, and the experimental results show that itoutperforms relevant baselines with lower communication costs.

AC2C: Adaptively Controlled Two-Hop Communication for Multi-Agent Reinforcement Learning