In 79 AD, the eruption of Mount Vesuvius buried Herculaneum, making the ancient city a frozen place. In the 18th century, a set of charred Herculaneum scrolls was excavated, which contained the Greek philosopher's profound reflections on pleasure. However, the broken state of these scrolls has puzzled researchers until recently, when three students from Egypt, Switzerland, and the United States successfully deciphered their contents through the collaborative efforts of AI technology, sparking a revolution in historical research.
This breakthrough in interpretation not only ignited a passion in the field of ancient studies, but also revealed the revolutionary impact that AI technology could have on our understanding of the ancient world. In addition to Herculaneum's scrolls, AI technology can also be used to study materials in mummified packages, revealing the lives of ordinary ancient Egyptians. The widespread use of this technique has opened new doors in the study of antiquity, allowing us to gain a deeper understanding of past civilizations.
The study of the Herculaneum scrolls is not unique, and hundreds of scrolls waiting to be identified have become an invaluable resource for researchers to explore. In the Vesuvius Challenge, three students won the grand prize of $700,000 by identifying more than 2,000 Greek letters on the scroll, equivalent to 5% of the scroll's content. This achievement broke with the pattern of slow progress in ancient studies, and scholars generally believe that it would lead to an exponential increase in the discovery of ancient Greek and Roman poetry, drama, and philosophical works.
Some of these texts may completely rewrite the history of key periods in the ancient world. Robert Fowler, president of the Herculaneum Society, said, "This is the society from which the modern Western world originated. "The intervention of AI has provided researchers with unprecedented tools, accelerated the interpretation of ancient texts, and provided us with a clearer view of our past.
The value of the Herculaneum scrolls lies not only in their contents, but also in the libraries of the ancient Roman era in which they are preserved. This luxurious Roman villa is believed to have belonged to Caesar's father-in-law, Senator Lucius Calpurnius Piso Cassoninus, while its library is the only surviving one from ancient Roman times. These carbonized scrolls have become a precious heritage of ancient research, and the application of AI technology has opened up new possibilities for the unlocking of this heritage.
In 1752, archaeological excavations in Herculaneum first uncovered about 1,800 scrolls, but due to the erosion of time, some of them were destroyed and became fragments. The state of preservation of these scrolls has always been a headache for researchers, and the consequence of directly unfolding the scrolls is that they become sparse. For years, scientists searched for ways to open these ancient scrolls, until the team of computer scientist Brent Sears at the University of Kentucky in the United States came up with a new idea of opening the scrolls digitally.
The software developed by the Sears team uses 3D computed tomography (CT) images to "virtually open" the rolled papyrus so that the surface image of the scroll is clearly visible. This innovation opened up new possibilities for interpreting the scroll's contents, however, the enormity of the project led Sears to look for more efficient ways to interpret them. Fortunately, the advice and support of Silicon Valley entrepreneur Nat Friedman gave new impetus to the challenge and laid the groundwork for what would become the Vesuvius Challenge.
In this challenge, young Luke Farrito won the grand prize by using machine learning algorithms to identify the faint textures on the reels and discover the ancient Greek word for "porphyras", which means "purple". Behind this discovery was a team effort, as did Yusuf Nader, a PhD student in Egypt, and Julian Sielig, a student at ETH Zurich. The efforts and innovations of these young scholars have injected fresh blood into the field of ancient studies and promoted the deepening of human understanding of history.
And that's not the end, Herculaneum's research continues. Friedman announced the 2024 "Vesuvius Challenge," encouraging researchers to decipher at least 90 percent of the characters of the four scanned scrolls. The competition will push the boundaries of machine learning to fully automate the tracking of the surface of each roll of papyrus and improve ink detection on the most damaged parts. This innovative effort will open the door to more reels for us and further deepen our understanding of Herculaneum.
The results of this study not only have a profound impact on ancient Greek philosophy, but also provide a paradigm for the application of machine learning techniques to the study of other types of hidden texts. In addition to Herculaneum's scrolls, machine learning techniques can be used to study papyri in mummified packages, revealing the lives of ordinary ancient Egyptians. The wide application of this technology will lead us into the past life, reveal the veil of history, and let us feel the style of ancient civilizations more truly.
However, this finding has also sparked debate about whether Herculaneum Villas warrants further investigation. Fowler and Richard Janko, a scholar of ancient papyrus manuscripts at the University of Michigan, are convinced that the villa's main library has never been discovered, and that thousands of scrolls may have been buried underground. This controversy will push ancient researchers to dig deeper into the mysteries of Herculaneum in an effort to restore the culture and knowledge of that era.
Taken together, Herculaneum's journey of AI interpretation not only lights up the study of antiquity, but also opens up new possibilities for the application of AI technology in historical research. This historical journey is like opening a door to the past, allowing us to see the glory of ancient civilizations more clearly, and also stimulating deeper thinking about history. With the help of machine learning, we may be able to understand the past more comprehensively and better understand the course of human civilization.