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

In recent years, on-policy reinforcement learning (RL) has been successfullyapplied to many different continuous control tasks. While RL algorithms areoften conceptually simple, their state-of-the-art implementations take numerouslow- and high-level design decisions that strongly affect the performance ofthe resulting agents. Those choices are usually not extensively discussed inthe literature, leading to discrepancy between published descriptions ofalgorithms and their implementations. This makes it hard to attribute progressin RL and slows down overall progress [Engstrom'20]. As a step towards fillingthat gap, we implement >50 such ``choices'' in a unified on-policy RLframework, allowing us to investigate their impact in a large-scale empiricalstudy. We train over 250'000 agents in five continuous control environments ofdifferent complexity and provide insights and practical recommendations foron-policy training of RL agents.

What Matters In On-Policy Reinforcement Learning? A Large-Scale Empirical Study