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 Tools annual

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 monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

We study the problem of obtaining accurate policy gradient estimates using afinite number of samples. Monte-Carlo methods have been the default choice forpolicy gradient estimation, despite suffering from high variance in thegradient estimates. On the other hand, more sample efficient alternatives likeBayesian quadrature methods have received little attention due to their highcomputational complexity. In this work, we propose deep Bayesian quadraturepolicy gradient (DBQPG), a computationally efficient high-dimensionalgeneralization of Bayesian quadrature, for policy gradient estimation. We showthat DBQPG can substitute Monte-Carlo estimation in policy gradient methods,and demonstrate its effectiveness on a set of continuous control benchmarks. Incomparison to Monte-Carlo estimation, DBQPG provides (i) more accurate gradientestimates with a significantly lower variance, (ii) a consistent improvement inthe sample complexity and average return for several deep policy gradientalgorithms, and, (iii) the uncertainty in gradient estimation that can beincorporated to further improve the performance.

Deep Bayesian Quadrature Policy Optimization