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

Sim-to-real transfer trains RL agents in the simulated environments and thendeploys them in the real world. Sim-to-real transfer has been widely used inpractice because it is often cheaper, safer and much faster to collect samplesin simulation than in the real world. Despite the empirical success of thesim-to-real transfer, its theoretical foundation is much less understood. Inthis paper, we study the sim-to-real transfer in continuous domain with partialobservations, where the simulated environments and real-world environments aremodeled by linear quadratic Gaussian (LQG) systems. We show that a popularrobust adversarial training algorithm is capable of learning a policy from thesimulated environment that is competitive to the optimal policy in thereal-world environment. To achieve our results, we design a new algorithm forinfinite-horizon average-cost LQGs and establish a regret bound that depends onthe intrinsic complexity of the model class. Our algorithm crucially relies ona novel history clipping scheme, which might be of independent interest.

Provable Sim-to-real Transfer in Continuous Domain with Partial Observations