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

Reinforcement learning (RL) techniques, while often powerful, can suffer fromslow learning speeds, particularly in high dimensional spaces. Decomposition oftasks into a hierarchical structure holds the potential to significantly speedup learning, generalization, and transfer learning. However, the current taskdecomposition techniques often rely on high-level knowledge provided by anexpert (e.g. using dynamic Bayesian networks) to extract a hierarchical taskstructure; which is not necessarily available in autonomous systems. In thispaper, we propose a novel method based on Sequential Association Rule Miningthat can extract Hierarchical Structure of Tasks in Reinforcement Learning(SARM-HSTRL) in an autonomous manner for both Markov decision processes (MDPs)and factored MDPs. The proposed method leverages association rule mining todiscover the causal and temporal relationships among states in differenttrajectories, and extracts a task hierarchy that captures these relationshipsamong sub-goals as termination conditions of different sub-tasks. We prove thatthe extracted hierarchical policy offers a hierarchically optimal policy inMDPs and factored MDPs. It should be noted that SARM-HSTRL extracts thishierarchical optimal policy without having dynamic Bayesian networks inscenarios with a single task trajectory and also with multiple tasks'trajectories. Furthermore, it has been theoretically and empirically shown thatthe extracted hierarchical task structure is consistent with trajectories andprovides the most efficient, reliable, and compact structure under appropriateassumptions. The numerical results compare the performance of the proposedSARM-HSTRL method with conventional HRL algorithms in terms of the accuracy indetecting the sub-goals, the validity of the extracted hierarchies, and thespeed of learning in several testbeds.

Autonomous Extraction of a Hierarchical Structure of Tasks in Reinforcement Learning, A Sequential Associate Rule Mining Approach