In an innovative fusion of technology and human expertise, scientists are leveraging artificial intelligence to unlock the mysteries of ancient scrolls carbonized by the infamous eruption of Mount Vesuvius in 79 A.D. These long-lost documents, discovered in the mid-18th century, were part of an extravagant villa library in Herculaneum, a town simultaneously doomed with Pompeii by the volcanic catastrophe.
The valuable Villa of the Papyri library holds the potential to profoundly expand our understanding of ancient philosophical thought, provided the scrolls can be deciphered. Until now, these relics have remained unreadable, reduced to fragile, charred blocks that crumble when physically unrolled due to the intense heat and volcanic ash they endured.
For over two and a half centuries, scholars across the globe have tried to make sense of these scrolls. Currently, the majority of these historical artifacts reside in the National Library of Naples. However, a modern twist in this age-old quest has arrived with the Vesuvius Challenge, launched in 2023 by tech entrepreneurs. This competition incentivizes breakthroughs in deciphering the scrolls through techniques like machine learning, computer vision, and geometry, promising monetary rewards for advancements.
Recently, participants in this challenge marked a momentous achievement. They successfully produced the inaugural image of the interior of one scroll among three housed at Oxford University's Bodleian Library. This groundbreaking progress was made possible by the expertise of Brent Seales, a computer scientist from the University of Kentucky and co-founder of the challenge, who rejoiced in the successful imaging of this scroll. He highlighted that the current discovery boasts more retrievable text than ever found in scanned Herculaneum scrolls.
The scroll underwent imaging using Diamond Light Source's cutting-edge technology in Harwell, near Oxford. Utilizing a particle accelerator, known as a synchrotron, the lab generated exceptionally powerful X-rays to scan the scroll. The resultant images were then meticulously combined through the power of AI, which enhanced text clarity and identified ink traces indicating preserved writing. This process culminated in a 3D representation, allowing scholars to virtually unroll and read the scroll using advanced segmentation techniques.
Despite these advances, the journey to fully decipher these ancient texts is just beginning. Presently, the AI has uncovered only limited portions of text, such as a few Greek phrases, including the word for disgust. Scholars are now encouraged to join the effort and help decode the remaining words.
Peter Toth, Cornelia Starks Curator of Greek Collections at the Bodleian, expressed optimism about ongoing improvements. We are still at the beginning of a long process. We need better images, and they remain very hopeful that image quality and text legibility will continue to get better, he told The Associated Press.
In the near future, there's potential for this innovative technology to be made portable, enabling further analyses without relocating the fragile scrolls to the lab. There might be something that can be moved, Toth suggested. This is critical, as around a thousand more scrolls await study in Naples, potentially harboring untold insights into the ancient world.
As scientists and AI work hand-in-hand, the promise of revealing long-silent voices from history seems closer than ever. This collaboration illustrates not only the power of technology to solve historical enigmas but also underscores humanity's enduring quest for knowledge.
