The Vesuvius Challenge is an academic competition to see who can use machine learning to translate ancient Roman scrolls recovered from the lost cities of Herculaneum and Pompeii. Since its excavation in 1752, researchers have been itching to find out what was inside the scroll, and now they have some answers.
Luke Farriter, a college student and former SpaceX summer intern, determined that one of the scrolls had those words written on it. Porphyrus,In other words, purple. Youssef Nader, an Egyptian biorobotics graduate student, later verified the discovery and took second place.
Both students were inspired by research previously conducted by Casey Handmer, who first discovered evidence of the presence of ink on scrolls.
For this achievement, Mr. Faritor and Mr. Nader received the First Letter Award. This prize gives him a prize of $50,000 if he finds 10 legible letters within an area of ​​4 square centimeters on the scroll. Luke gets his $40,000 and Yusef gets his $10,000.
(Credit: ScrollPrize)
This is interesting for several reasons. The scrolls were buried and carbonized by the eruption of Mount Vesuvius in 79 AD. Despite being excavated hundreds of years ago, very few scrolls have never been read, as most of them shatter when opened. Although several attempts were made in the years following its discovery, no one had attempted to open the scroll since the 19th century. The contents of the scroll left behind are a complete mystery.
Dr. Brent Shields and his team took high-resolution X-rays of the scrolls and digitally flattened them using virtual opening techniques. There was just one problem. The ink used by the ancient Romans does not show up on X-rays. The researchers correctly deduced that machine learning could tell the difference between ink and blank papyrus. Hence his Vasuvius Challenge, a competition to see who could load a scroll into a computer first, was born.
Editor’s picks
Mr. Farriter and Mr. Nader are the first to make award-winning contributions to this challenge. Fariter used Handmer’s existing research in conjunction with machine learning to identify the ink portion of the X-ray. As the model trained, more and more letters started appearing, and eventually the first word was found. Meanwhile, Nader focused on ink detection on his own as part of his Kraggle Ink Detection Award, which is also part of the Vesuvius Challenge. Nader’s method, as well as Farritor’s, are available on his GitHub.
With this discovery, researchers believe it is only a matter of time before the team deciphers the entire scroll and wins the Vesuvius Challenge. In fact, Nader’s model has already produced new images showing even more characters that until this week he had been hidden from human eyes for 2,000 years.
Get our best stories!
Apply New features now Get the top stories delivered to your inbox every morning.
This newsletter may contain advertisements, deals or affiliate links. By subscribing to our newsletter, you agree to our Terms of Use and Privacy Policy. You can unsubscribe from the newsletter at any time.