Closing the Gap

in Non-Latin-Script Data