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

Q-learning suffers from overestimation bias, because it approximates themaximum action value using the maximum estimated action value. Algorithms havebeen proposed to reduce overestimation bias, but we lack an understanding ofhow bias interacts with performance, and the extent to which existingalgorithms mitigate bias. In this paper, we 1) highlight that the effect ofoverestimation bias on learning efficiency is environment-dependent; 2) proposea generalization of Q-learning, called \emph{Maxmin Q-learning}, which providesa parameter to flexibly control bias; 3) show theoretically that there exists aparameter choice for Maxmin Q-learning that leads to unbiased estimation with alower approximation variance than Q-learning; and 4) prove the convergence ofour algorithm in the tabular case, as well as convergence of several previousQ-learning variants, using a novel Generalized Q-learning framework. Weempirically verify that our algorithm better controls estimation bias in toyenvironments, and that it achieves superior performance on several benchmarkproblems.

Maxmin Q-learning: Controlling the Estimation Bias of Q-learning