Scientists Use Artificial Intelligence to Decipher Ancient Herculaneum Scrolls
In a groundbreaking achievement, a team of young scientists has successfully deciphered passages from previously unreadable scrolls found in the ancient city of Herculaneum. This achievement marks a significant milestone in the study of ancient history and has been made possible through the use of artificial intelligence (AI) technology.
Over 2,000 years ago, Mount Vesuvius erupted, burying the city of Herculaneum and its library, which contained more than a thousand scrolls. In the 1800s, these scrolls were discovered, but attempts to read them resulted in their destruction due to their delicate, carbonized state.
However, Youssef Nader, Luke Farritor, and Julian Schilliger managed to overcome this challenge by utilizing machine learning, a form of AI, to unveil the secrets hidden within the ancient Greek texts. By applying AI technology, the team was able to decode four passages from the scrolls without ever unrolling them. Their success has earned them the prestigious grand prize of $700,000 (€650,000) in the Vesuvius Challenge.
Through this groundbreaking work, the team has shed new light on the writings of Philodemus, the philosopher-in-residence at the villa where the scrolls were found. By delving into the texts, researchers hope to gain a deeper understanding of the classical world and the philosophy of living a good life, as described by Philodemus.
The process of deciphering the scrolls involved the use of computed tomography (CT) scans, which created detailed 3D images of the rolled papyrus. The AI technology then helped identify regions of ink in the flattened segments of the papyrus. By combining and reconstructing the individual spots of ink, the AI algorithm was able to unveil the hidden text.
The Vesuvius Challenge, organized by Brent Seales of the University of Kentucky in collaboration with Nat Friedman and Daniel Gross, proved to be a significant catalyst for progress in deciphering the Herculaneum scrolls. It was student and SpaceX intern Luke Farritor who made a crucial breakthrough by training a machine learning model to detect patterns in the cracked texture of the ancient Greek text. This discovery led to the decipherment of the word ΠΟΡΦΥΡΑϹ, meaning purple.
Thanks to the contributions of Egyptian PhD student Youssef Nader and Swiss robotics student Julian Schilliger, further progress was made in reading columns of text and creating 3D maps of the papyrus scrolls, respectively.
Looking ahead, the next Vesuvius Challenge aims to read an entire work or scroll by the end of 2024. The discoveries made through AI technology offer exciting prospects for researchers to uncover more hidden texts and gain unparalleled insights into the ancient world.
The successful use of artificial intelligence in deciphering the Herculaneum scrolls not only highlights the valuable role that technology plays in historical research but also demonstrates the incredible power of AI to uncover secrets from the past. This groundbreaking achievement sets the stage for even more remarkable discoveries and opens a unique window into the classical world.
