DeepMind may serve cookies to analyse traffic to this site. Information about your use of this site is shared with DeepMind for that purposeSee detailsOK, got it

Restoring and attributing ancient texts using deep neural networks

Χαῖρε! Welcome to Ithaca's interactive interface. Please follow the instructions below to begin restoring and attributing ancient Greek inscriptions. You will also find more information on the Ithaca project, links to the article and examples of Ithaca in action.

Ancient History relies on disciplines such as Epigraphy, the study of inscribed texts known as “inscriptions”, for evidence of the thought, language, society and history of past civilizations. However, over the centuries many inscriptions have been damaged to the point of illegibility, transported far from their original location, and their date of writing is steeped in uncertainty. We present Ithaca, the first Deep Neural Network for the textual restoration, geographical and chronological attribution of ancient Greek inscriptions. Ithaca is designed to assist and expand the historian's workflow: its architecture focuses on collaboration, decision support, and interpretability.

Example of restored inscription
Restoration of a damaged inscription, recording a decree from 485/4 BCE concerning the Acropolis of Athens (IG I3 4B, CC BY-SA 3.0, WikiMedia).

While Ithaca alone achieves 62% accuracy when restoring damaged texts, as soon as historians use Ithaca their performance leaps from 25% to 72%, confirming this synergistic research aid's impact. Ithaca can attribute inscriptions to their original findspot with 71% accuracy and can date them with a distance of less than 30 years from ground-truth ranges, redating key texts of Classical Athens and contributing to topical debates in Ancient History.

Example results table
Ithaca's experimental results. Evaluated methods for text restoration, geographical (region) and chronological attribution (date) on I.PHI's test set. For CER and Years, lower scores are better (↓).

This work shows how models like Ithaca can unlock the cooperative potential between AI and historians, transformationally impacting the way we study and write about one of the most significant periods in human history.

Ithaca was conceived and researched by Yannis Assael*, Thea Sommerschield*, Brendan Shillingford, Mahyar Bordbar, John Pavlopoulos, Marita Chatzipanagiotou, Ion Androutsopoulos, Jonathan Prag, and Nando de Freitas. This web experience was developed and built on Google Cloud by Justin Grayston, Benjamin Maynard, and Ricardo Cardenas.

Use Ithaca for your research

Input your ancient Greek epigraphic text in the box below to attribute the text to its original place and time of writing, and restore any missing characters. Mark missing characters or spaces as dashes (-), and the ones be predicted as question marks (?). Each query can predict up to 10 question marks (consecutive or not). To predict longer character sequences or to inspect more restoration hypotheses, we refer researchers to the Colaboratory notebook.

0/760