DeepMind's new artificial intelligence model helps decipher ancient writings, find history and location.
Machine learning methods provide archaeologists with new tools that can help them understand the past - especially when it comes to deciphering ancient texts. The latest example is an artificial intelligence model developed by Alphabet's subsidiary DeepMind, which not only helps to restore missing text in ancient Greek writing, but also offers suggestions on when the text was written and its possible geographical origin.
Thea Sommerschield, a historian and machine learning expert who helped create the model, told reporters at a press briefing that "the writings are really important because they are a direct source of evidence ... written directly by the ancient people themselves."
It is really difficult for a person to use all the available, relevant information and discover the basic patterns. The new software, Ithaca, is taught on a database of approximately 78,608 ancient Greek scripts, each of which is labeled with metadata describing where and when it was written (as far as historians know). Like all machine learning systems, Ithaca searches for examples in this information, encodes this information in complex mathematical models, and uses these results to suggest text, history, and origin.
These statistics are encouraging, but we must remember that Ithaca cannot function independently of human experience. In fact, his proposals are based on information collected by traditional archeological methods