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

Reinforcement learning approaches have long appealed to the data managementcommunity due to their ability to learn to control dynamic behavior from rawsystem performance. Recent successes in combining deep neural networks withreinforcement learning have sparked significant new interest in this domain.However, practical solutions remain elusive due to large training datarequirements, algorithmic instability, and lack of standard tools. In thiswork, we introduce LIFT, an end-to-end software stack for applying deepreinforcement learning to data management tasks. While prior work hasfrequently explored applications in simulations, LIFT centers on utilizinghuman expertise to learn from demonstrations, thus lowering online trainingtimes. We further introduce TensorForce, a TensorFlow library for applied deepreinforcement learning exposing a unified declarative interface to common RLalgorithms, thus providing a backend to LIFT. We demonstrate the utility ofLIFT in two case studies in database compound indexing and resource managementin stream processing. Results show LIFT controllers initialized fromdemonstrations can outperform human baselines and heuristics across latencymetrics and space usage by up to 70%.

LIFT: Reinforcement Learning in Computer Systems by Learning From Demonstrations