Israeli Scientists Use AI to Reconstruct Damaged Babylonian Clay Tablets


Inscribed Babylonian clay tablet with cuneiform script

Recently, an Israeli research team from Ariel University used artificial intelligence to reconstruct fragmentary Akkadian clay tablets inscribed with cuneiform scripts. The team's paper "Restoration of Fragmentary Babylonian Texts Using Recurrent Neural Networks" was published in the Proceedings of the National Academy of Sciences.

Clay cuneiform tablets are main sources for deciphering information regarding ancient Mesopotamian history and culture. Many of these tablets are damaged, leading to missing information. In the article, the research team investigate the possibility of formulating the ancient language using recurrent neural networks, so that machines could be trained to help experts or independently restore lost Akkadian texts from Achaemenid period Babylonia. They use digitized texts to train advanced machines to learn algorithms and restore daily economic and administrative documents from the Persian empire (6th to 4th centuries B.C.). As the amount of digitized sample texts grow, the model can be trained to restore damaged texts belonging to other genres, such as scientific or literary texts. This study is a first step for a large-scale reconstruction of ancient texts.

Source: Xinhua News Agency

Photo source: British Museum