AI Bots to Decode Gaps and Interpret Ancient Languages

31 Mar 2022

Harvard researchers and historians, in collaboration, have created an artificial intelligence (AI) system that can read and interpret ancient scriptures.

In the experimental rounds, it was tried on the ancient Greek text, which showed 62% accuracy when it is used alone to fill gaps in inscriptions and 72% accuracy when the system's ideas were interpreted by a historian - over three times greater than when historians work alone, according to the findings.

About the Robot

The AI-based robot is named Ithaca, after the Greek island where King Odysseus resided.

The Ithaca system can detect word patterns and the age of texts ranging from imperial decrees to Sappho's poems.

Researchers have developed a system that can fill in the blanks in ancient Greek inscriptions to establish when and where they were written.

Dr. Thea Sommerschield from the University of Venice is one of the historians working on this AI-enabled technology. According to her, the inscriptions are significant as they provide a peek into the former civilizations' thoughts, languages, cultures, and history.

However, most parts of the inscriptions have deteriorated over time. As a result, the words are either incomplete or unintelligible. Relocation from their original site and ineffective radiocarbon dating procedures on stone are some of the reasons that have led to this deterioration.

How Does AI-Based Ithaca Work?

Over 63,000 transcribed ancient Greek manuscripts were fed into Ithaca, allowing it to spot patterns in the order of words and letters, as well as affiliations between words or phrases and the text's age and provenance.

Out of all the 63,000+ manuscripts, the system predicted the origins of around 7,811 inscriptions. It also offered a list of words and letters to fill intentionally manufactured gaps in the writings, sorted by probability.

According to the study conducted by the team of historians and researchers, Ithaca predicted the date of all inscriptions with a margin error of 30 years. It was able to recognize the source 71% of the time.

Applications of Ithaca

The technique may be applied to any media and any historical written language. The system can also be trained on Greek literary works written by famous authors. Then there is the possibility of developing AI systems that can provide information about text authorship.

Ithaca had been used on a set of rulings, most of which were discovered from Athens. It implies that some Athenian collections could be dated to 424 Before Christ instead of 448-447 Before Christ; it coincides with the latest dating techniques breakthrough.

Though it may be a minor distinction, this 30-year shift has significant implications for understanding classical Athens' political history. It allows better reflection on classical literary inscriptions.

The application of AI through Ithaca has the potential to expand the toolkit of academics to analyze historical records.

However, AI also faces the same constraints as historians do, which is facing inadequacies in the historical records. It is only effective as a tool to assist scholars in asking questions about specific evidence and making correlations.

Other Advancements in Language AI

Researchers at Massachusetts Institute of Technology’s Computer Science and Artificial Intelligence Laboratory (CSAIL) recently made a significant breakthrough in the field of AI language, demonstrating a new system that can automatically interpret a lost language without the need for extensive knowledge of its interconnections with other languages.

They also demonstrated that their algorithm could determine language relationships on its own. This was used to back up recent research suggesting that the ancient language of Europe and Iberian is not linked to Basque (Spain).

The main objective of the researchers is to make the system able to interpret lost languages that linguists have been unable to decipher for decades, using just a few thousand words.

In conclusion, AI bots are being equipped with all human aspects that could suitably benefit humans in various unexplored ways. Researchers are now inculcating the basic technology, AI, into different industries to get the best possible outcome, and it is working wonders.

Artificial intelligence and machine learning technologies have been adapted in various industries in innovative ways. The industries include food technology, education technology, space technology, automobile sector, agriculture technology, etc. AI has been growing rapidly in the agriculture technology sector.  


Request a Sample


Twitter Feeds