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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Transfer reinforcement learning (RL) aims at improving the learningefficiency of an agent by exploiting knowledge from other source agents trainedon relevant tasks. However, it remains challenging to transfer knowledgebetween different environmental dynamics without having access to the sourceenvironments. In this work, we explore a new challenge in transfer RL, whereonly a set of source policies collected under diverse unknown dynamics isavailable for learning a target task efficiently. To address this problem, theproposed approach, MULTI-source POLicy AggRegation (MULTIPOLAR), comprises twokey techniques. We learn to aggregate the actions provided by the sourcepolicies adaptively to maximize the target task performance. Meanwhile, welearn an auxiliary network that predicts residuals around the aggregatedactions, which ensures the target policy's expressiveness even when some of thesource policies perform poorly. We demonstrated the effectiveness of MULTIPOLARthrough an extensive experimental evaluation across six simulated environmentsranging from classic control problems to challenging robotics simulations,under both continuous and discrete action spaces. The demo videos and code areavailable on the project webpage: https://omron-sinicx.github.io/multipolar/.

MULTIPOLAR: Multi-Source Policy Aggregation for Transfer Reinforcement Learning between Diverse Environmental Dynamics