A team of researchers from Tel Aviv University and Ariel University has developed an artificial intelligence model that translates the text of Akkadian cuneiform tablets into English. They presented their work in a recent article entitled “Translating Akkadian to English with Neural Machine Translation” in the journal PNAS Nexus. The scholars used natural language processing methods (NLP) to automatically translate the Akkadian texts using two methods: converting the cuneiform Unicode glyphs directly to English and converting Unicode glyphs to a Latin transliteration first, then to English. Both methods achieved good results for sentences that were short or medium-length (118 characters or less), but the method using the Latin transliteration first was the most accurate. The AI model may speed up the translation of cuneiform texts and would be best used as part of a human-machine collaboration in which Assyriologists correct the initial computer translations.
OFF-SITE LINKS:
- https://www.jpost.com/archaeology/article-741982
- https://academic.oup.com/pnasnexus/article/2/5/pgad096/7147349
Read more BREAKING NEWS articles here: https://biblearchaeology.org/current-events-list
