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

Deep reinforcement learning (DRL) has gained great success by learningdirectly from high-dimensional sensory inputs, yet is notorious for the lack ofinterpretability. Interpretability of the subtasks is critical in hierarchicaldecision-making as it increases the transparency of black-box-style DRLapproach and helps the RL practitioners to understand the high-level behaviorof the system better. In this paper, we introduce symbolic planning into DRLand propose a framework of Symbolic Deep Reinforcement Learning (SDRL) that canhandle both high-dimensional sensory inputs and symbolic planning. Thetask-level interpretability is enabled by relating symbolic actions tooptions.This framework features a planner -- controller -- meta-controllerarchitecture, which takes charge of subtask scheduling, data-driven subtasklearning, and subtask evaluation, respectively. The three componentscross-fertilize each other and eventually converge to an optimal symbolic planalong with the learned subtasks, bringing together the advantages of long-termplanning capability with symbolic knowledge and end-to-end reinforcementlearning directly from a high-dimensional sensory input. Experimental resultsvalidate the interpretability of subtasks, along with improved data efficiencycompared with state-of-the-art approaches.

SDRL: Interpretable and Data-efficient Deep Reinforcement Learning Leveraging Symbolic Planning