The eruption of Vesuvius in 79 AD. destroyed Pompeii and nearby Herculaneum, where hot gases burned hundreds of ancient papyri in the library of a huge luxurious villa.

Researchers have launched an international competition to read charred papyri after demonstrating that an artificial intelligence program can distinguish between letters and symbols in high-resolution x-ray images of fragile, unfolded documents.

Scientists led by Professor Brent Sills, a computer scientist at the University of Kentucky, have been able to read the ink on the surface and underlying layers of papyrus by training a machine learning algorithm to detect subtle differences in papyrus texture captured in photographs. X-ray scans

“We have shown how to read the ink (ss on papyrus) of Herculaneum. This gives us the opportunity to reveal 50%, 70%, maybe 80% of the entire collection,” Sills calculated. “We built it vessel. Now we want everyone to get on board and sail away.”

For the Vesuvius test, Sills’ team shared software and thousands of 3D X-ray images of two rolled papyri and three sections of papyrus. The scientists hope that the $250,000 prize pool will attract research teams from around the world who will improve the artificial intelligence system and accelerate the deciphering of the only discovered library of antiquity, “frozen” in time, without a deficit.

“We’re running a competition to improve our ability to extract more and more text,” says Sills.