Artificial intelligence's uncanny ability to spot patterns in vast amounts of data could finally unravel some of the ancient world's thorniest mysteries.
Researchers working with companies like IBM and Google's Deepmind are on the verge of deciphering ancient texts once thought unreadable — and even “cracking” an unknown language from nearly two millennia before the birth of Christ.
AI allows researchers to scroll through images much faster than humans, and the techniques could answer fundamental questions about the history of the language and potentially uncover lost works by Greek and Roman writers.
Cracking an unknown language
The 1,400 tablets in Linear A date from 1800 BC and have never been deciphered despite many efforts
A mysterious unknown language, 'Linear A', discovered on tablets in Crete in 1900, has never been deciphered – but AI might be able to crack the code.
Among the world's most famous examples of unknown languages are stones and tablets written in the foreign language 'LINEar A'. These are considered the main script used by the Minoan civilization, a Bronze Age kingdom led by King Minos.
The 1,400 tablets in Linear A date from 1800 BC and, despite many efforts, have never been deciphered; another, newer script found on tablets on the island, Linear B, was cracked in 1953.
The breakthrough came thanks to researchers who realized that repeated words in the language could be place names on the island and that Linear B could be similar to Ancient Greek.
The 1,400 tablets in Linear A date from 1800 BC and, despite many efforts, have never been deciphered
But Linear A poses a much more difficult challenge, because the 'precursor language' is completely unknown – and no text in the language has ever been translated.
Arthur Evans, the researcher who found the tablets, said Linear A is one of the earliest writings discovered and its similarity to images shows its essential place in the evolution of human language.
Researchers from MIT and Google's AI lab Deepmind have used AI to automatically translate texts in Linear B – the first time this has been done, raising hopes that AI could one day unravel texts in Linear A.
Several researchers use techniques such as data mining and natural language processing to unravel patterns within Linear AI.
A team from MIT and Deepmind is working on a new system that can decipher lost languages by examining the relationships between different languages.
Reading an old library
Artificial intelligence could finally give researchers a way to read passages from the scrolls, which have remained undeciphered for nearly two millennia
The Herculaneum Scrolls were discovered in ruins near Pompeii – astonishingly fragile charred scrolls preserved in the aftermath of the volcanic eruption in 76 AD.
Herculaneum was buried under boiling mud during the eruption and the scrolls were found in 1750 in a luxurious villa thought to have belonged to Julius Caesar's father-in-law.
Unlike other libraries of the period, which had dissolved due to contact with the air, the scrolls were preserved, but until now unreadable.
Artificial intelligence could finally give researchers a way to read passages from the scrolls, which remained undeciphered for nearly two millennia after the eruption.
Research led by University of Kentucky professor Brent Seales and with a cash prize offered by Silicon Valley investors uses AI models to decipher the scrolls.
The Herculaneum Scrolls were discovered in ruins near Pompeii – astonishingly fragile charred scrolls preserved in the aftermath of the volcanic eruption in 76 AD.
Computer science students used computed tomography (similar to X-ray scans) and artificial intelligence to “read” marks on a papyrus scroll.
The Herculaneum Scrolls are the only surviving library from classical antiquity, and researchers hope there may be undiscovered works by poets and playwrights from ancient Greece and Rome.
Described as a 'potential treasure trove', a computer science student unlocked the first word – 'purple' – from one scroll to win $40,000 in the Vesuvius Challenge.
'This word is our first delve into an unopened old book, reminiscent of royalty, wealth and even mockery. What will the context show? Pliny the Elder explores 'purple' in his 'natural history' as a production process for Tyrian purple from shellfish,” Professor Seales said.
'The Gospel of Mark describes how Jesus was mocked for wearing a purple robe before the crucifixion.
“What this specific scroll discusses is still unknown, but I believe it will be revealed soon.”
The Mystery of the Nazca Lines
The Nazca Lines, which show animals and strange humanoids including one described as 'an astronaut', were first found in 1927 – but no one has explained their purpose.
Conspiracy theorists suggest that the vast lines spanning the Nazca Plateau may have been created by aliens.
Now AI drives the process of discovering new Nazca lines
Now AI drives the process of discovering new Nazca lines
Scientists suggest that they were most likely used as routes for processions, and that they are so large that they are visible to the gods.
Now AI is fueling the process of discovering new Nazca lineages, dating back to 500 B.C. and 500 AD. B.C. have been carved into the landscape of Peru by humans.
Researchers at Yamagata University have now accelerated the process of discovering new Nazca Lines by 21 times using deep learning in collaboration with IBM – and are now using AI to search for 'geoglyphs' missed in previous searches.
Now AI drives the process of discovering new Nazca lines
The researchers are racing against time as erosion and climate change endanger the geoglyphs.
The researchers write: 'It is imperative to identify and protect as many geoglyphs as possible.
Now that they can discover more of the mysterious geoglyphs, researchers hope to understand more about how and why they were created.