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

A critical challenge in multi-agent reinforcement learning(MARL) is formultiple agents to efficiently accomplish complex, long-horizon tasks. Theagents often have difficulties in cooperating on common goals, dividing complextasks, and planning through several stages to make progress. We propose toaddress these challenges by guiding agents with programs designed forparallelization, since programs as a representation contain rich structural andsemantic information, and are widely used as abstractions for long-horizontasks. Specifically, we introduce Efficient Multi-Agent Reinforcement Learningwith Parallel Program Guidance(E-MAPP), a novel framework that leveragesparallel programs to guide multiple agents to efficiently accomplish goals thatrequire planning over $10+$ stages. E-MAPP integrates the structuralinformation from a parallel program, promotes the cooperative behaviorsgrounded in program semantics, and improves the time efficiency via a taskallocator. We conduct extensive experiments on a series of challenging,long-horizon cooperative tasks in the Overcooked environment. Results show thatE-MAPP outperforms strong baselines in terms of the completion rate, timeefficiency, and zero-shot generalization ability by a large margin.

E-MAPP: Efficient Multi-Agent Reinforcement Learning with Parallel Program Guidance