AI Restores 3000-Year-Old Babylonian Poem

Researchers from LMU Munich and Baghdad University have used artificial intelligence to reconstruct a poem titled "The Babylonian Hymn." It was lost over 2000 years ago, writes Decrypt.

The text praises Babylon and the god Marduk. It was written 3000 years ago, and it was last studied in 100 BC.

The team behind the project reported that the work was created from 30 clay fragments unearthed over many years, and AI was used to piece the work together.

"We used a specialized artificial intelligence-based program to analyze and match text fragments based on combinations of cuneiform signs," said Enrique Jimenez, a professor of ancient Near Eastern languages at LMU Munich.

The expert, along with colleagues, uses methods based on natural language processing to determine to which text the fragments belong.

Working with the Electronic Babylonian Library platform, which contains 1402 manuscripts, researchers apply n-gram matching as the primary method of reconstruction. Other approaches are also used, such as lexical composition overlap and the search for the longest common sequences of text.

Jimenez emphasized that the discovered poem was significant enough for contemporaries and was included in the Babylon curriculum.

In an article for the journal Iraq, he and co-author Anmar Fadhil suggested that the creator of the poem likely belonged to the priestly class of Babylon, as in one section of the poem the priests are described as "free citizens."

The anthem praises the natural resources and beauties of the city, respect for foreigners, and support for the poor.

"They do not humiliate foreigners among them. They protect the humble and support the weak. Under their care, the poor and the dispossessed can thrive. They grant orphans comfort and favor," the composition states.

Jimenez noted that artificial intelligence is becoming indispensable for researchers, "especially for restoring damaged or fragmented texts."

"Although languages such as Akkadian and Sumerian are still underrepresented in LLM, we are actively working on improving computational tools for the study of the ancient Near East," noted the expert.

Earlier, researchers from DeepMind and the Ca' Foscari University of Venice developed the AI algorithm Ithaca to restore damaged ancient texts.

Let us remind you that in October 2021, specialists from Google AI recreated the destroyed paintings of Gustav Klimt using artificial intelligence.