Recent advances in Apertium, a free/open-source rule-based machine translation platform for low-resource languages

Abstract This paper presents an overview of Apertium, a free and open-source rule-based machine translation platform. Translation in Apertium happens through pipeline modular tools, the platform continues to be improved as more language pairs are added. Several advances have been implemented since last publication, including some new optional modules: module that allows rules process recursive structures at structural transfer stage, deals with contiguous discontiguous multi-word expressions, resolves anaphora aid translation. Also highlighted is hybridisation statistical modules augment pipeline, methods existing modules. includes morphological disambiguation, weighted transfer, lexical selection learn from limited data. The also discusses how like can critical part access technology for so-called low-resource languages, which might ignored or deemed unapproachable by popular corpus-based technologies. Finally, released unreleased pairs, concluding brief look supplementary tools prove valuable users well developers. All Apertium-related code, data, free/open-source available https://github.com/apertium .