AI Book Translation 01.26.18

AI Translation Interprets Mysterious Book

For more than a century, academics, cryptographers, and conspiracy theorists have debated and theorized about the meaning of the Voynich manuscript, an ornate tome written in an unknown language and encrypted with an obscure code. With the cipher proving unbreakable, theories about the document’s purpose have run wild, with hypotheses ranging from an alchemist’s guide to an alien’s diary. However, AI translation technology recently made some groundbreaking discoveries about the book’s language and possible meaning.

Greg Kondrak, a computer scientist in the AI lab at the University of Alberta, said he employed statistical algorithms that perform AI translation to help unlock the mystery of the Voynich manuscript. The algorithms’ first discovery was the book likely was written in Hebrew, according to CTVNews.

The scientist used algorithms that have demonstrated a high degree of accuracy in identifying languages. These algorithms boasted a 97 percent success rate when translating the United Nations Universal Declaration of Human Rights into 380 languages.

With the language identified, Kondrak and his AI translation technology took a try at breaking the code. Using the same processes employed for language identification, the algorithms determined that the book’s author disguised the meaning of words by shuffling the order of their letters and omitting their vowels.

The algorithms revealed that the manuscript’s first complete sentence read, “She made recommendations to the priest, man of the house and me and people.” The text also includes the words “farmer,” “light,” “air” and “fire,” according to CTVNews.

The combination of the book’s words and the illustrations may provide some hints about the manuscript’s meaning. The book’s drawings include multiple depictions of nude women, alongside images of stars and planets, astrological signs and various types of plants. This could indicate that the book’s topic relates to women’s health issues.

Kondrak’s work could represent the first phase of the resolution of the Voynich mystery, an enigma that has long puzzled scholars. The manuscript was written in the early 1400s and was named after Polish book dealer Wilfrid Voynich, who bought the work in 1912. In the intervening years, a range of prominent experts tried and failed to interpret the book, including Britain’s Bletchley Park cryptographers, who successfully cracked the Nazis’ supposedly unbreakable Enigma code during World War II.

AI translation and natural language processing technology increasingly are finding ways to break codes that have proven unsolvable using conventional methods. For example, USC professor Kevin Knight has been using the CARMEL supercomputer to solve some the most vexing cyphers in human history.

Knight recently put CARMEL on the case of the Zodiac serial killer, leveraging the technology to decipher coded letters authored by the criminal. After more than 40 years of failed decoding attempts, Carmel was able to solve a significant portion of the Zodiac’s most arcane code, possibly leading to clues about the killer’s identity, according to Fox News.

Nirel Marofsky is project analyst for the cognitive engine and application ecosystem at Veritone. She acts as a liaison to strategic partners, integrating developers and their capabilities into the Veritone Platform. Learn more about our platform and join the Veritone developer ecosystem today.