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

Deep Reinforcement Learning (DRL) has shown promise in various networkingenvironments. However, these environments present several fundamentalchallenges for standard DRL techniques. They are difficult to explore andexhibit high levels of noise and uncertainty. Although these challengescomplicate the training process, we find that in practice we can substantiallymitigate their effects and even achieve state-of-the-art real-world performanceby addressing a factor that has been previously overlooked: the skewed inputtrace distribution in DRL training datasets.  We introduce a generalized framework, Plume, to automatically identify andbalance the skew using a three-stage process. First, we identify the criticalfeatures that determine the behavior of the traces. Second, we classify thetraces into clusters. Finally, we prioritize the salient clusters to improvethe overall performance of the controller. Plume seamlessly works across DRLalgorithms, without requiring any changes to the DRL workflow. We evaluatedPlume on three networking environments, including Adaptive Bitrate Streaming,Congestion Control, and Load Balancing. Plume offers superior performance inboth simulation and real-world settings, across different controllers and DRLalgorithms. For example, our novel ABR controller, Gelato trained with Plumeconsistently outperforms prior state-of-the-art controllers on the livestreaming platform Puffer for over a year. It is the first controller on theplatform to deliver statistically significant improvements in both videoquality and stalling, decreasing stalls by as much as 75%.

Plume: A Framework for High Performance Deep RL Network Controllers via Prioritized Trace Sampling