Скачать или смотреть AI Deciphers hidden patterns in ancient texts and IT'S WILD!

I have a Patreon page!
  / themetatron  

DeepMind's Aeneas represents a groundbreaking advancement in the field of digital archaeology and ancient text analysis, specifically designed to help historians decode and contextualize ancient Roman inscriptions. Named after the wandering hero of Graeco-Roman mythology, Aeneas builds upon Ithaca, DeepMind's earlier work using AI to restore, date and place ancient Greek inscriptions. This multimodal generative neural network takes an inscription's text and image as input, using a transformer-based decoder to process textual information while specialized networks handle character restoration and dating using text, with geographical attribution also utilizing images of the inscriptions.

The AI system was trained on the Latin Epigraphic Dataset (LED), comprising over 176,000 Latin inscriptions from across the ancient Roman world, which researchers curated by cleaning, harmonizing and linking records from decades of work by historians in creating digital collections. Aeneas demonstrates remarkable capabilities, restoring damaged inscriptions with a Top-20 accuracy of 73% in gaps of up to ten characters, which only decreases to 58% when the restoration length is unknown. The system can attribute an inscription to one of 62 ancient Roman provinces with 72% accuracy and can date texts within 13 years of the date ranges provided by historians.

What makes these Latin inscriptions particularly valuable to historians is that they offer first-hand evidence of ancient thought, language, society and history written by ancient people themselves across all social classes, not just history written by the elite. Approximately 1,500 new Latin inscriptions are discovered every year, reflecting how commonplace these chunks of engraved text were in the Roman world, from imperial decrees to street graffiti, providing insights into both important historical developments and everyday life. When tested with 23 historians, Aeneas helped spur research ideas for 90% of inscriptions and led to more accurate determinations of where and when the inscriptions originated, with the best results occurring when historians used the AI model together with their skills as researchers rather than relying solely on one or the other.

The significance of Aeneas extends beyond mere text restoration - it fundamentally changes how historians can approach the vast corpus of Roman epigraphy. According to leading Cambridge classics professor Mary Beard, the process of identifying and contextualizing newly discovered inscriptions can often amount to "guesswork" or following a "hunch," but Aeneas can rapidly scan its wide-ranging database containing thousands of Latin inscriptions, automatically surfacing useful parallels. Rather than seeking to automate epigraphy entirely, the DeepMind team is interested in crafting a tool that will integrate with the workflow of a historian, giving researchers trying to analyze a specific inscription many hypotheses to work from, saving them the effort of sifting through records by hand.

#translation #ancientlanguages #hotnews

artificial intelligence, DeepMind, ancient Rome, Latin inscriptions, epigraphy, machine learning, historical research, text restoration, archaeological technology, classical studies, digital humanities, neural networks, multimodal AI, Roman Empire, ancient history