AI that helps historians connect the past Open Source • History Books • Artificial Intelligence 6 143 Framer Launch websites with enterprise needs at startup speeds. Design Tools • Website Builder • Artificial Intelligence
Aeneas is a new open-source AI model from Google DeepMind that helps historians restore, date, and place fragmentary ancient inscriptions. It analyzes both text and images to provide new context on ancient texts.
Aeneas places inscriptions within 13 years of historian-provided date ranges and attributes texts to one of 62 ancient Roman provinces with 72% accuracy using multimodal analysis of text and images. This accelerates the traditionally manual process of contextualizing inscriptions within their historical period and location.
Aeneas restores gaps in damaged texts with 73% top-20 accuracy for gaps up to 10 characters and 58% accuracy for unknown-length gaps. It identifies textual parallels across 176,000+ Latin inscriptions, helping historians interpret isolated fragments and piece together meaning from weathered or deliberately defaced texts.
Aeneas is freely available at predictingthepast.com for researchers, students, educators, and museum professionals. Code and the Latin Epigraphic Dataset (176,000+ inscriptions) are open-sourced, enabling widespread adoption and collaborative research without licensing fees.
Aeneas produces probabilistic date distributions rather than single fixed dates, capturing multiple scholarly hypotheses quantitatively. For example, it modeled competing theories about the Res Gestae Divi Augusti by showing two distinct peaks in its dating prediction, grounding historical debates in linguistic and contextual data.
Aeneas searches across thousands of Latin inscriptions in seconds, retrieving parallels that share similarities in wording, syntax, formulas, or provenance. It converts each inscription into a historical fingerprint using embeddings, enabling historians to situate fragments within broader historical context far faster than manual comparison.