Israeli Experts Leverage AI To Translate Ancient Cuneiform Texts

An A.I. model created by researchers from Tel Aviv University and Ariel University in Israel provides an automated English translation for cuneiform Akkadian text. Translations are done quickly and accurately by these universities’ new model.

Preserving and disseminating the cultural heritage of Ancient Mesopotamia has taken an important step forward, say experts. This is a significant move towards safeguarding this historic tradition for future generations.

Specialists in Assyriology, a field involving archaeological, historical, cultural, and linguistic investigations of ancient Mesopotamia and its civilization called Assyria, use cuneiform to make sense of the Akkadian texts, one of the oldest written forms. Such an effort has been ongoing for many years.

Cuneiform, from the word “wedge-shaped,” gets its name due to how it was written. In ancient times, people used a reed stylus to make wedge-shaped marks on clay tablets, forming the cuneiform writing system.

Tel Aviv and Ariel researchers have created an Artificial Intelligence model that can effortlessly convert cuneiform translations of Akkadian texts into English. This will save much effort that would have otherwise gone into manual conversions.

The Jerusalem Post recently reported an interesting piece of news according to the Biblical Archaeology educational project – a mosaic discovered in an ancient Jewish house in northern Israel.

After their god, Ashshur, Assyria was located in the Mesopotamian plain. In 721 B.C., the Assyrian army marched from the north and brought down the Northern Kingdom of Israel, also abducting its ten tribes, who were then lost to history.

Hundreds of ancient clay tablets dating back to 3400 B.C., written in cuneiform and discovered by archaeologists, have been found in Mesopotamia.

Tablets which contain written records have been difficult to interpret due to the limited number of experts knowledgeable enough to read them. Despite this challenge, there is a vast amount of tablets that remain untranslated.

On May 2, 2023, a team of researchers from Ariel University and Tel Aviv University, led by Dr. Shai Gordin and Dr. Gai Gutherz, published their research in the PNAS Nexus journal titled “Translating Akkadian to English with neural machine translation.”

The researchers designed two versions of the new machine learning model for automatic translation from Akkadian: one translating Latinized inscriptions into English and the other translating cuneiform Unicode elements directly.

This study tested two versions of a machine translation model. The one using Latin transliteration produced better results, resulting in a score of 37.47 in the BLEU4 evaluation – a method to measure how accurately machine and human translations match.

The program has its most success when sentences of 118 characters or less are translated; however, for some input, there may be “hallucinations” – syntactically correct but inaccurate English.

Using translation as a primary text processing is a feasible solution. Consequently, researchers recommend using machine translation in combination with human effort so that scientists can rectify and improve the results generated by models.

Hundreds of thousands of clay tablets written in cuneiform preserve evidence of ancient Mesopotamia’s political, social, economic, and scientific history. These documents offer insight into this ancient era’s cultural dynamics and activities.

“but most of these documents remain untranslated and inaccessible due to their sheer number and the limited number of experts who can read them.”

Translation has long been a fundamental part of human activity and is rooted in the science of writing. It often necessitates expertise in the translated language and its related cultural context to convey meaning effectively. This makes for a complex process involving various considerations.

Technological innovations, including optical character recognition and machine translation, are permitting more digital tools to be designed for translating ancient languages. Yet the difficulties in this area are still considerable.

To fully comprehend these ancient texts, it is necessary to understand a language that has been extinct for a long time. Furthermore, these documents are often found in poor condition and can be incomplete.

The Ukrainian Union Conference news site originally posted this story. It has since been revised and restructured while maintaining the same content.

The successful implementation of AI in translating ancient cuneiform marks a significant milestone in studying ancient civilizations. This technological breakthrough has expedited decipherment, providing researchers with knowledge and insights into the past.

As AI continues to evolve, we can expect further advancements in the field of historical linguistics, enabling us to unlock the secrets of our ancestors and preserve our shared heritage for generations to come.

Source: Adventist News Network

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top