Machine Translation of Texts from Languages with Low Digital Resources: A Systematic Review

This research conducted a systematic review of related works on machine translation of languages ​​with low digital resources. First, we carried out the information search in the databases: ScienceDirect, IEEE Xplore, ACM Digital Library. Eighteen articles were collected following inclusion and exclusion criteria, considering a search period from 2016 to 2022. Subsequently, we analyzed and classified these articles according to the libraries developed and/or used based on machine learning, statistics, or grammar. The results indicate that pre-training and morphological segmentation techniques with finite state machines and machine learning techniques improve the translation of languages ​​with low digital resources. In addition, according to the articles compiled in the specialized databases, in Mexico, unlike other countries that we analyzed, there are few publications on the translation of languages ​​with low digital resources, and we mostly found research papers published in international conferences.