Details
- Google DeepMind introduced Aeneas, the first AI model designed to contextualize ancient Roman inscriptions, in a Nature paper published July 23, 2025.
- The project was developed with the University of Nottingham and in partnership with researchers at the Universities of Warwick, Oxford, and Athens University of Economics and Business.
- Aeneas uses multimodal analysis—combining text and images—to restore missing text, date inscriptions within 13 years, and determine their geographic origins with 72% accuracy, drawing from a dataset of 176,000 Latin inscriptions.
- Building on DeepMind's previous Ithaca system for Greek, Aeneas is tailored to support historians' workflows and interpret fragmented ancient texts.
- An interactive version is available for free to the public, and both the code and dataset have been open-sourced to encourage wider research.
Impact
Aeneas marks a leap forward in historical research, speeding the analysis of ancient texts and deepening connections across fragmented sources. By open-sourcing its resources and fostering partnerships between AI experts and historians, DeepMind shapes a new model for collaborative discovery. The technology's adaptability points to future breakthroughs across languages and archaeological mediums.