To try Greek Room, register at app.greekroom.org now!
The Greek Room is a project to improve the quality and efficiency of Bible translation with software tools that help translators and quality control consultants with spell-checking, terminology consistency, translation drafting and other tasks.
Thousands of languages around the world still lack Bible translations; most of them have extremely few resources such as extensive lexicons, existing translations, spell checkers or even spelling standards. Such ultra-low resource languages present challenges for natural language processing tools.
Wildebeest
This tool checks a text for a wide range of character level problems such as encoding errors, characters in the wrong script, characters in non-canonical forms and spurious control characters. The Wildebeest analysis script provides feedback on likely issues in a text. The Wildebeest normalization script corrects problems that can be fixed with very high confidence. See the Wildebeest section for more information, including examples from applying Wildebeest to the 1009 translations in the eBible corpus.
Spell Checking
The Greek Room offers a spell checker based on phonetic string similarity and word alignment, rather than relying on the need of laboriously curated lists of words and morphological rules used in traditional spell checkers. It allows to identify spelling inconsistencies early on in Bible translation projects. Click here to see some examples, including on the English NRSV.
Word Alignment
Word alignment supports applications such as spell-checking and learning word order and morphology. Even more directly, it indicates the strength of translation consistency. With its color-coded word alignment strengths, the Greek Room specifically helps to pinpoint words that are unusual translations, or missing or spurious words. Hovering with the mouse over a word in the source or target language will reveal additional information such as the gloss of a word. Click here to see some examples, including English-German and English-Hindi.
About
Research team: Dr. Ulf Hermjakob and Joel Mathew; University of Southern California (USC) Information Sciences Institute (ISI)
Greek Room in the news:
USC Viterbi,
Washington Post,
Catholic New Agency,
ABC News,
BBC Radio 4 (UK),
BBC Brazil