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

Training an agent to solve control tasks directly from high-dimensionalimages with model-free reinforcement learning (RL) has proven difficult. Apromising approach is to learn a latent representation together with thecontrol policy. However, fitting a high-capacity encoder using a scarce rewardsignal is sample inefficient and leads to poor performance. Prior work hasshown that auxiliary losses, such as image reconstruction, can aid efficientrepresentation learning. However, incorporating reconstruction loss into anoff-policy learning algorithm often leads to training instability. We explorethe underlying reasons and identify variational autoencoders, used by previousinvestigations, as the cause of the divergence. Following these findings, wepropose effective techniques to improve training stability. This results in asimple approach capable of matching state-of-the-art model-free and model-basedalgorithms on MuJoCo control tasks. Furthermore, our approach demonstratesrobustness to observational noise, surpassing existing approaches in thissetting. Code, results, and videos are anonymously available athttps://sites.google.com/view/sac-ae/home.

Improving Sample Efficiency in Model-Free Reinforcement Learning from Images