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

In this article, we propose a novel algorithm for deep reinforcement learningnamed Expert Q-learning. Expert Q-learning is inspired by Dueling Q-learningand aims at incorporating semi-supervised learning into reinforcement learningthrough splitting Q-values into state values and action advantages. We requirethat an offline expert assesses the value of a state in a coarse manner usingthree discrete values. An expert network is designed in addition to theQ-network, which updates each time following the regular offline minibatchupdate whenever the expert example buffer is not empty. Using the board gameOthello, we compare our algorithm with the baseline Q-learning algorithm, whichis a combination of Double Q-learning and Dueling Q-learning. Our results showthat Expert Q-learning is indeed useful and more resistant to theoverestimation bias. The baseline Q-learning algorithm exhibits unstable andsuboptimal behavior in non-deterministic settings, whereas Expert Q-learningdemonstrates more robust performance with higher scores, illustrating that ouralgorithm is indeed suitable to integrate state values from expert examplesinto Q-learning.

Expert Q-learning: Deep Reinforcement Learning with Coarse State Values from Offline Expert Examples