Language Log

« previous post | next post »

If you've ever had any doubt about the positive potential of AI for fundamental linguistic research of various types, here's a powerful example that will set your mind at rest.

"First passages of rolled-up Herculaneum scroll revealed:  Researchers used artificial intelligence to decipher the text of 2,000-year-old charred papyrus scripts, unveiling musings on music and capers."  By Jo Marchant, Nature (2/5/24).

doi: https://doi.org/10.1038/d41586-024-00346-8

With four striking illustrations, including a video and an animation, plus a separate related visual showing how the feat was accomplished.

A team of student researchers has made a giant contribution to solving one of the biggest mysteries in archaeology by revealing the content of Greek writing inside a charred scroll buried 2,000 years ago by the eruption of Mount Vesuvius. The winners of a contest called the Vesuvius Challenge trained their machine-learning algorithms on scans of the rolled-up papyrus, unveiling a previously unknown philosophical work that discusses senses and pleasure. The feat paves the way for artificial intelligence (AI) techniques to decipher the rest of the scrolls in their entirety, which researchers say could have revolutionary implications for our understanding of the ancient world.

The achievement has ignited the usually slow-moving world of ancient studies. It’s “what I always thought was a pipe dream coming true”, says Kenneth Lapatin, curator of antiquities at the J. Paul Getty Museum in Los Angeles, California, who was not involved in the contest. The revealed text discusses sources of pleasure including music, the taste of capers and the colour purple. “It’s an historic moment,” says classicist Bob Fowler at the University of Bristol, UK, one of the prize judges. The three students, from Egypt, Switzerland and the United States, who revealed the text share a US$700,000 grand prize.

The scroll is one of hundreds of intact papyri excavated in the eighteenth century from a luxury Roman villa in Herculaneum, Italy. These lumps of carbonized ash — known as the Herculaneum scrolls — are the only library that survives from the ancient world, but are too fragile to open.

The winning entry, announced on 5 February, reveals hundreds of words across more than 15 columns of text, corresponding to around 5% of an entire scroll. “The contest has cleared the air on all the people saying will this even work,” says Brent Seales, a computer scientist at the University of Kentucky, Lexington, and co-founder of the prize. “Nobody doubts that anymore.”

…

In the centuries after the scrolls were discovered, many people have attempted to open them, destroying some and leaving others in pieces. Papyrologists are still working to decipher and stitch together the resulting, horribly fragmented, texts. But the chunks in the worst condition — the most hopeless cases, adding up to perhaps 280 entire scrolls — were left intact. They’re held mostly in the National Library of Naples, Italy, with a few in Paris, London and Oxford, UK.

Seales has been trying to read these concealed texts for nearly 20 years. His team developed software to “virtually unwrap” the surfaces of rolled-up papyri using three-dimensional computed tomography (CT) images. In 2019, he carried two of the scrolls from the Institut de France in Paris to the Diamond Light Source particle accelerator near Oxford to make high-resolution scans.

Mapping the surfaces was time consuming, however, and the carbon-based ink used to write the scrolls has the same density as papyrus in CT scans, so it was impossible to differentiate in imaging. Seales and his colleagues wondered whether machine-learning models might be trained to ‘unwrap’ the scrolls and distinguish the ink. But making sense of all the data was a gigantic task for his small team.

Seales was approached by Silicon Valley entrepreneur Nat Friedman, who had become intrigued by the Herculaneum scrolls after watching a talk by Seales online. Friedman suggested opening the challenge to contestants. He donated $125,000 to launch the effort and raised hundreds of thousands more on Twitter, and Seales released his software along with the high-resolution scans. The team launched the Vesuvius Challenge in March 2023, setting a grand prize for reading 4 passages, of at least 140 characters each, before the end of the year.

…

The colour purple

A key innovation came in the middle of last year, when US entrepreneur and former physicist Casey Handmer noticed a faint texture in the scans, similar to cracked mud — he called it “crackle” — that seemed to form the shapes of Greek letters. Luke Farritor, an undergraduate studying computer science at the University of Nebraska-Lincoln, used the crackle to train a machine-learning algorithm, revealing the word porphyras, ‘purple’, which won him the prize for unveiling the first letters in late October. An Egyptian PhD student in Berlin, Youssef Nader, who followed with even clearer images of the text, came second.

…

Their code was released with less than three months for contestants to scale up their reads before the deadline for the final prize of 31 December. “We were biting our nails,” says Friedman. But in the final week, the competition received 18 submissions. A technical jury checked entrants’ code, then passed 12 submissions to a committee of papyrologists who transcribed the text and assessed each entry for legibility. Only one fully met the prize criteria: a team formed by Farritor and Nader, along with Julian Schilliger, a Swiss robotics student at the Swiss Federal Institute of Technology Zurich.

Music, pleasure and capers

The content of most of the previously opened Herculaneum scrolls relates to the Epicurean school of philosophy, and seems to have formed the working library of a follower of the Athenian philosopher Epicurus, who lived from 341 to 270 BC, named Philodemus. The new text doesn’t name the author but, from a rough first read, say Fowler and Nicolardi, it is probably also by Philodemus. As well as pleasurable tastes and sights, it includes a figure called Xenophantus, possibly a flute-player of that name mentioned by the ancient authors Seneca and Plutarch, whose evocative playing apparently caused Alexander the Great to reach for his weapons.

Lapatin says the topics discussed by Philodemus and Epicurus are still relevant. “The basic questions Epicurus was asking are the ones that face us all as humans. How do we live a good life? How do we avoid pain?” But “the real gains are still ahead of us”, he says. “What’s so exciting to me is less what this scroll says, but that the decipherment of this scroll bodes well for the decipherment of the hundreds of scrolls that we had previously given up on.”

….

This amazing project shows how the cooperative efforts of brilliant human minds, combined with the massive, raw, calculatory power of AI can unlock secrets that have been sealed up for millennia.

 

Selected readings

• "ytš ḥṭ ḏ lqml śʿ[r w]zqt" (11/10/22)
• "Canaanite in the news again" (2/2/23)
• "Decipherment of Linear A" (5/24/23)
• "The Origin(s) of Writing" (3/19/22) — Phaistos disc; Linear A / B
• "Whence cometh linguistic meaning?" (6/14/18) — Rosetta stone; Phaistos Disc; Linear A; Indus Valley script
• "Toward the decipherment of Harappan" (2/14/22)
• "Decipherment of Linear Elamite" (9/5/22)
• "Solving Linear A" (11/29/19)
• "Minoan language Linear A linked to Linear B in new research" (5/15/22)
• "Translation as cryptography as translation" (11/19/12)
• "Hype over AI and Classical Chinese / Literary Sinitic" (11/9/23)

[Thanks to jjtk and Gene Hill]

February 6, 2024 @ 5:45 pm · Filed by Victor Mair under Artificial intelligence, Decipherment, Language and archeology

Permalink