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 this work, we propose to apply trust region optimization to deepreinforcement learning using a recently proposed Kronecker-factoredapproximation to the curvature. We extend the framework of natural policygradient and propose to optimize both the actor and the critic usingKronecker-factored approximate curvature (K-FAC) with trust region; hence wecall our method Actor Critic using Kronecker-Factored Trust Region (ACKTR). Tothe best of our knowledge, this is the first scalable trust region naturalgradient method for actor-critic methods. It is also a method that learnsnon-trivial tasks in continuous control as well as discrete control policiesdirectly from raw pixel inputs. We tested our approach across discrete domainsin Atari games as well as continuous domains in the MuJoCo environment. Withthe proposed methods, we are able to achieve higher rewards and a 2- to 3-foldimprovement in sample efficiency on average, compared to previousstate-of-the-art on-policy actor-critic methods. Code is available athttps://github.com/openai/baselines

Scalable trust-region method for deep reinforcement learning using Kronecker-factored approximation