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

Puzzles and board games represent several importantclasses of AI problems, but also represent difficultcomplexity classes. In this paper, we propose a deep learningbased alternative to train a neural network model to find solutionstates of the popular puzzle game Sokoban. The network trainsagainst a classical solver that uses theorem proving as the oracleof valid and invalid games states, in a setup that is similar tothe popular adversarial training framework. Using our approach,we have been able to verify the validity of a Sokoban puzzle upto an accuracy of 99% on the test set. We have also been ableto train our network to generate the next possible state of thepuzzle board up to an accuracy of 99% on the validation set. Wehope that through this approach, a trained neural network willbe able to replace human experts and classical rule-based AI ingenerating new instances and solutions for such games.

Generation of Sokoban Stages using Recurrent Neural Networks