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

Modelling and exploiting teammates' policies in cooperative multi-agentsystems have long been an interest and also a big challenge for thereinforcement learning (RL) community. The interest lies in the fact that ifthe agent knows the teammates' policies, it can adjust its own policyaccordingly to arrive at proper cooperations; while the challenge is that theagents' policies are changing continuously due to they are learningconcurrently, which imposes difficulty to model the dynamic policies ofteammates accurately. In this paper, we present \emph{ATTention Multi-AgentDeep Deterministic Policy Gradient} (ATT-MADDPG) to address this challenge.ATT-MADDPG extends DDPG, a single-agent actor-critic RL method, with twospecial designs. First, in order to model the teammates' policies, the agentshould get access to the observations and actions of teammates. ATT-MADDPGadopts a centralized critic to collect such information. Second, to model theteammates' policies using the collected information in an effective way,ATT-MADDPG enhances the centralized critic with an attention mechanism. Thisattention mechanism introduces a special structure to explicitly model thedynamic joint policy of teammates, making sure that the collected informationcan be processed efficiently. We evaluate ATT-MADDPG on both benchmark tasksand the real-world packet routing tasks. Experimental results show that it notonly outperforms the state-of-the-art RL-based methods and rule-based methodsby a large margin, but also achieves better performance in terms of scalabilityand robustness.

Modelling the Dynamic Joint Policy of Teammates with Attention Multi-agent DDPG