A computer-science student, Luke Farritor, has successfully used a machine-learning algorithm to read text inside a carbonized scroll from the ancient city of Herculaneum. The scroll had been unreadable since its preservation by a volcanic eruption in AD 79. This groundbreaking achievement could unlock texts from the only surviving library from Greco-Roman antiquity. Farritor’s technology discerned Greek letters within the scroll by detecting minute variations in surface texture.
 Â
Key Points:Â
- Luke Farritor developed an algorithm that identifies Greek letters on the papyrus of a carbonized Herculaneum scroll, making previously unreadable ancient texts accessible.
- The Vesuvius Challenge offers monetary awards for deciphering texts from these ancient scrolls. Farritor won the ‘first letters’ prize of $40,000.
- The scrolls, discovered in the 18th century, could potentially revolutionize the understanding of ancient history and literature as they contain unique works not found elsewhere.
- The machine-learning technique capitalizes on subtle texture differences to distinguish inked areas on the papyrus, shedding light on an “invisible library” of undecipherable ancient texts.
- The wider application of AI in the study of ancient texts is growing, with tools being developed to interpret and date various ancient inscriptions and languages.
