DeepMind’s latest AI model

Machine learning techniques are giving archaeologists new tools to help them understand the past, especially when it comes to deciphering ancient texts. The most recent example is an AI model developed by Alphabet subsidiary DeepMind that not only helps in restoring missing text from ancient Greek inscriptions but also suggests when the text was written (within 30 years) and its likely geographic origins.

A historian and machine learning expert – Thea Sommerschield who assisted in the model’s development, told journalists at a press conference that Inscriptions are certainly important since they are the direct origin of evidence written directly by ancient people themselves.

These texts are frequently damaged due to their age, making restoration a worthwhile challenge. Because they are frequently inscribed on inorganic materials such as stone or metal, methods such as radiocarbon dating cannot be used to determine when they were written.

Sommerschield, who co-led the research with DeepMind staff researcher Yannis Assael stated that for solving these tasks, epigraphers look for textual and contextual parallels in identical inscriptions. However, it is extremely difficult for a human to harness all available, relevant data and uncover underlying patterns.

Machine learning can be utilized in these scenarios.

Ithaca is a new piece of software that was trained on a dataset of 78,608 ancient Greek inscriptions, each of which is labeled with metadata describing where and when it was written (to the best of historians’ knowledge). Ithaca, like all machine learning systems, looks for patterns in this data, encodes it in complex mathematical models, and then utilizes these inferences for suggesting text, date, and origins.

The scientists who developed Ithaca claim that it is 62 percent accurate when restoring letters in damaged texts in a paper published in Nature. It can attribute an inscription’s geographic origins to one of 84 ancient world regions with 71 percent accuracy and can date a text to within 30 years of its known year of writing on average.

These are encouraging statistics, but it’s important to remember that Ithaca cannot function without human assistance. Its recommendations are ultimately based on data gathered through traditional archaeological methods, and its creators are positioning it as merely another tool in a larger set of forensic methods, rather than a fully-automated AI historian. Ithaca was created as a supplement to help historians, Sommerschield explained.

Ithaca, according to Eleanor Dickey, a professor of classics at the University of Reading who specializes in ancient Greek and Latin sociolinguists, is an exciting development that may enhance our knowledge of the ancient world. She did, however, add that a 62 percent accuracy rate for restoring lost text was not reassuring — when people count on it, they will need to remember that it is wrong about one-third of the time — and that she was unsure how the software would fit into current academic methodologies.

DeepMind, for instance, highlighted tests that showed the model improved historians’ accuracy in restoring missing text in ancient inscriptions from 25% to 72%. Dickey, however, points out that those being tested were students, not professional epigraphers. She claims that while AI models are widely available, they cannot or should not replace the small cadre of specialized academics who decipher texts.

It is not yet clear to what level utilization of this tool by fairly qualified editors would result in an enhancement in the editions that are currently available — but it will be fascinating to find out, Dickey said. She went on to say that she wanted to try out the Ithaca model for herself. The software, as well as its open-source code, is accessible for anyone to test online.

Ithaca and it’s antecedent (named Pythia and released in 2019) have already been utilized for aiding recent archaeological debates, including assisting in the dating of inscriptions discovered in Athens’ Acropolis. However, the software’s true potential has yet to be realized.

Sommerschield emphasizes that the true value of Ithaca may be found in its adaptability. Although it was trained on ancient Greek inscriptions, it was easily adaptable to other ancient scripts. “The architecture of Ithaca makes it truly applicable to any ancient language, not just Latin, but Mayan, cuneiform; any written medium — papyri, manuscripts, she explained. There are numerous opportunities.

Source link