Skip to main content

New research uses deep learning to restore ancient Greek texts


Yannis Assael, and colleagues from the department of classics have been involved in an innovative interdisciplinary research project, 'Restoring ancient text using deep learning: a case study on Greek epigraphy'.

This work is a collaboration between our department, the faculty of Classics (ancient history) and DeepMind. Inscriptions are one of the main direct sources of new evidence from the ancient world, but the majority have suffered damage over the centuries, and parts of the text are illegible or lost. Restoring the missing or damaged text is one of the main undertakings of the discipline of Epigraphy.

The research team trained Pythia, the first ancient text restoration model that recovers missing characters from a damaged text input using deep neural networks. Bringing together the disciplines of ancient history and deep learning, this work offers a fully automated aid to the text restoration task, providing ancient historians with multiple textual restorations, as well as the confidence level for each hypothesis.

The combination of machine learning and epigraphy has the potential to impact meaningfully the study of inscribed texts, and widen the scope of the historian's work. For this reason, an online Python notebook, Pythia, and PHI-ML’s processing pipeline have been open sourced on GitHub ( By so doing, it is the authors' hope to aid future research and inspire further interdisciplinary work.

Read the full story at DeepMind's blog post:

read more:
This work has been accepted by EMNLP 2019 (