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 paper, a new offline actor-critic learning algorithm is introduced:Sampled Policy Gradient (SPG). SPG samples in the action space to calculate anapproximated policy gradient by using the critic to evaluate the samples. Thissampling allows SPG to search the action-Q-value space more globally thandeterministic policy gradient (DPG), enabling it to theoretically avoid morelocal optima. SPG is compared to Q-learning and the actor-critic algorithmsCACLA and DPG in a pellet collection task and a self play environment in thegame Agar.io. The online game Agar.io has become massively popular on theinternet due to intuitive game design and the ability to instantly competeagainst players around the world. From the point of view of artificialintelligence this game is also very intriguing: The game has a continuous inputand action space and allows to have diverse agents with complex strategiescompete against each other. The experimental results show that Q-Learning andCACLA outperform a pre-programmed greedy bot in the pellet collection task, butall algorithms fail to outperform this bot in a fighting scenario. The SPGalgorithm is analyzed to have great extendability through offline explorationand it matches DPG in performance even in its basic form without extensivesampling.

Sampled Policy Gradient for Learning to Play the Game Agar.io