DeepMind's AI Enhances Restoration of Ancient Greek Inscriptions

DeepMind, in collaboration with the University of Oxford, has developed Ithaca, an advanced AI system designed to assist historians in restoring damaged ancient Greek inscriptions. This deep neural network not only predicts missing text but also estimates the original location and dating of these artifacts. Ithaca's performance has been evaluated using a dataset of nearly 8,000 inscriptions, demonstrating a 62% accuracy in text restoration. When historians collaborate with Ithaca, their restoration accuracy improves significantly, reaching 72%. Additionally, Ithaca can attribute inscriptions to their original location with 71% accuracy and date them within 30 years of their true creation date. This collaborative approach between AI and historians offers a promising tool for advancing the study of ancient texts.