The Bible, the weapon to improve algorithms

It contains more than 31,000 verses that scientists used to produce more than 1.5 million unique combinations of verses

The tools to translate texts on the internet are counted by hundreds and translate phrases from English to Spanish, through Chinese or Japanese, among others. Now, researchers have been inspired by the Bible to go one step further in the translation of texts.

A group of experts has seen in the Bible "a large set of aligned parallel text data without previously exploding". The sacred text features 31,000 verses, each version, a few sentences that have served scientists to produce more than 1.5 million unique combinations of source and target verses for machine learning training sets.

To mark the style for the study, researchers refer to the length of the sentence, the use of passive or active voices and the selection of words that could result in texts with varying degrees of simplicity or formality.

The team used 34 versions of different styles, such as the "King James Version" to the "Basic English Bible," according to the research. The texts were incorporated into two algorithms: a statistical machine translation system called "Moses" and a neural network framework commonly used in machine translation, "Seq2Seq".

According to research published in the journal Royal Society Open Science, this is not the first parallel data set created for the translation of styles, but it is the first that uses the Bible. "The English Bible comes in many different written styles, so it is the perfect source text to work on the translation of styles."

Although different versions of the Bible were used to train the computer code, systems that translate the style of any written text for different audiences could ultimately be developed.