TL;DR:
- Israeli researchers develop an AI model for translating the cuneiform Akkadian text into English
- Assyriology specialists have struggled to understand ancient Akkadian texts written in cuneiform
- Tel Aviv and Ariel Universities’ AI model simplifies the translation process
- The program has limitations, occasionally producing accurate but contextually incorrect translations
- Machine translation can be used in collaboration with human experts for improved results
- Hundreds of thousands of clay tablets written in cuneiform hold untapped historical information
- The majority of these tablets remain untranslated due to limited expert resources
- The translation is a complex process requiring knowledge of languages and cultural contexts
- Digital tools like optical character recognition and machine translation aid the translation process
- Ancient languages present unique challenges, including fragmented texts and extinct linguistic communities.
Main AI News:
Tel Aviv and Ariel Universities in Israel have made significant strides in the preservation and dissemination of Ancient Mesopotamia’s cultural heritage through the development of an artificial intelligence (AI) model. This AI model is capable of automatically translating the cuneiform Akkadian text into English, a breakthrough achievement according to experts.
For years, Assyriology specialists, who delve into archaeological, historical, cultural, and linguistic studies of Assyria and ancient Mesopotamia, have devoted themselves to comprehending Akkadian texts inscribed in cuneiform—a form of writing dating back centuries. Cuneiform, meaning “wedge-shaped,” originated from the practice of making wedge-shaped imprints on clay tablets using a reed stylus.
The arduous efforts of these experts may now be alleviated, thanks to the groundbreaking work carried out by researchers from Tel Aviv and Ariel Universities. They have developed an AI model capable of automatically translating Akkadian text written in cuneiform into English.
This remarkable development was reported by the esteemed educational project “Biblical Archaeology,” with The Jerusalem Post as its source. The implications of this breakthrough extend far beyond simple translation, as it represents a crucial step in safeguarding and disseminating the historical and cultural significance of Ancient Mesopotamia.
Assyria, named after the deity Ashshur, the foremost god in the Assyrian pantheon, once thrived in the Mesopotamian plain. In a pivotal moment in history, in 721 BC, the Assyrian army surged from the north, conquering the Northern Kingdom of Israel and capturing ten tribes, effectively rendering them lost to the annals of time.
Throughout ancient Mesopotamia, archaeologists have unearthed hundreds of thousands of clay tablets inscribed in cuneiform, with origins dating back as early as 3400 BC. However, the translation of these tablets has proven to be an immense challenge, given the limited number of experts capable of deciphering their contents.
Dr. Shai Gordin from Ariel University, along with Dr. Gai Gutherz and their colleagues from Tel Aviv University, recently published their groundbreaking findings in the esteemed PNAS Nexus journal. Titled “Translating Akkadian to English with neural machine translation,” their research sheds light on the significant advancements achieved in this field (Gai Gutherz, Shai Gordin, Luis Sáenz, Omer Levy, Jonathan Berant. PNAS Nexus, Volume 2, Issue 5, May 2023, pgad096).
The researchers devised a novel machine-learning model, presenting two variations of an automatic translation from Akkadian. One version employs Latin transliteration, while the other directly translates cuneiform Unicode elements into English. Surprisingly, the former demonstrated more promising outcomes, garnering a notable score of 37.47 in the Best Bilingual Evaluation Understudy 4 (BLEU4). This metric measures the level of correspondence between machine and human translations of identical text, highlighting the impressive accuracy achieved by the AI model.
The effectiveness of the program for translating sentences of 118 characters or less has been highlighted. However, it is worth noting that in certain cases, the program generated syntactically correct but inaccurate English translations, which can be considered “hallucinations.”
According to Dr. Gordin, the translation can serve as a primary text-processing tool in many instances. The authors of the study propose the utilization of machine translation as a part of human-machine collaboration, where human scientists can rectify and enhance the output generated by the models.
The significance of the clay tablets, numbering in the hundreds of thousands, inscribed in cuneiform cannot be overstated. These tablets offer valuable insights into the political, social, economic, and scientific history of ancient Mesopotamia. However, the vast majority of these documents have remained untranslated and inaccessible due to their sheer volume and the limited number of experts capable of deciphering them.
The researchers conclude that translation is an inherently human activity that has a long-standing scientific history since the advent of writing. It is a multifaceted process that demands expertise in two distinct languages, along with an understanding of different cultural contexts.
The increasing prevalence of digital tools aimed at facilitating translation is driven by advancements in fields like optical character recognition and machine translation. However, ancient languages pose a formidable challenge in this regard. Deciphering and comprehending these languages necessitates knowledge of long-extinct linguistic communities, and the texts themselves are often highly fragmented.
Conlcusion:
The development of an artificial intelligence model capable of automatically translating the cuneiform Akkadian text into English holds significant implications for the market. This breakthrough technology has the potential to revolutionize the field of Assyriology and the study of ancient Mesopotamia by making a vast amount of untapped historical information more accessible and understandable. With the ability to decipher and translate ancient texts more efficiently, researchers, historians, and cultural enthusiasts can gain deeper insights into the political, social, economic, and scientific history of ancient civilizations.
This advancement in machine translation technology also paves the way for potential commercial applications, such as developing language learning tools, facilitating cross-cultural research collaborations, and supporting cultural heritage preservation initiatives. Overall, this innovation opens up new opportunities in the market for AI-driven language translation solutions and demonstrates the immense potential of advanced technologies in unlocking the secrets of our past.
