# Restoring ancient text using deep learning: a case study on Greek epigraphy

arXiv: https://arxiv.org/abs/1910.06262 · HTML: https://mlsanity.com/paper/1910.06262/

## Key facts

- Cited by 25 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 14 Oct 2019.

## Built on by (most cited first, 25 of 25)

- [InCoder: A Generative Model for Code Infilling and Synthesis](https://mlsanity.com/paper/2204.05999/) (2022)
- [AGTGAN: Unpaired Image Translation for Photographic Ancient Character Generation](https://mlsanity.com/paper/2303.07012/) (2023)
- [Restoring and Mining the Records of the Joseon Dynasty via Neural Language Modeling and Machine Translation](https://mlsanity.com/paper/2104.05964/) (2021)
- [Filling the Gaps in Ancient Akkadian Texts: A Masked Language Modelling Approach](https://mlsanity.com/paper/2109.04513/) (2021)
- [NLP for The Greek Language: A Longer Survey](https://mlsanity.com/paper/2408.10962/) (2024)
- [HistRED: A Historical Document-Level Relation Extraction Dataset](https://mlsanity.com/paper/2307.04285/) (2023)
- [BERT in Plutarch's Shadows](https://mlsanity.com/paper/2211.05673/) (2022)
- [Ancient Script Image Recognition and Processing: A Review](https://mlsanity.com/paper/2506.19208/) (2025)
- [DeepScribe: Localization and Classification of Elamite Cuneiform Signs Via Deep Learning](https://mlsanity.com/paper/2306.01268/) (2023)
- [Reviving Cultural Heritage: A Novel Approach for Comprehensive Historical Document Restoration](https://mlsanity.com/paper/2507.05108/) (2025)
- [Logion: Machine Learning for Greek Philology](https://mlsanity.com/paper/2305.01099/) (2023)
- [Insightful analysis of historical sources at scales beyond human capabilities using unsupervised Machine Learning and XAI](https://mlsanity.com/paper/2310.09091/) (2023)
- [Instruct-Tuning Pretrained Causal Language Models for Ancient Greek Papyrology and Epigraphy](https://mlsanity.com/paper/2409.13870/) (2024)
- [Deep Aramaic: Towards a Synthetic Data Paradigm Enabling Machine Learning in Epigraphy](https://mlsanity.com/paper/2310.07310/) (2023)
- [EpiAgent: An Agent-Centric System for Ancient Inscription Restoration](https://mlsanity.com/paper/2604.09367/) (2026)
- [Restoring Ancient Ideograph: A Multimodal Multitask Neural Network Approach](https://mlsanity.com/paper/2403.06682/) (2024)
- [Review of Computational Epigraphy](https://mlsanity.com/paper/2406.06570/) (2024)
- [Lacuna Language Learning: Leveraging RNNs for Ranked Text Completion in Digitized Coptic Manuscripts](https://mlsanity.com/paper/2407.12247/) (2024)
- [Automating Violence Detection and Categorization from Ancient Texts](https://mlsanity.com/paper/2503.08192/) (2025)
- [LakotaBERT: A Transformer-based Model for Low Resource Lakota Language](https://mlsanity.com/paper/2503.18212/) (2025)
- [MESA: A Training-Free Multi-Exemplar Deep Framework for Restoring Ancient Inscription Textures](https://mlsanity.com/paper/2604.17390/) (2026)
- [Leveraging External Knowledge for Historical Document Restoration via Retrieval-Augmented Large Language Models](https://mlsanity.com/paper/2607.21936/) (2026)
- [Stoicheia: Character-Level Masked Diffusion for Ancient Greek Textual Restoration, Parsing, and Metrical Scansion](https://mlsanity.com/paper/2608.07249/) (2026)
- [Text Restoration of Ancient Documents with Language Models](https://mlsanity.com/paper/2608.28170/) (2026)
- [Apollo Restore: A Foundation LLM for Historical Greek Optimized for Fill-in-the-Middle Restoration of Ancient Greek Texts](https://mlsanity.com/paper/2609.22455/) (2026)

## Abstract (arXiv)

Ancient history relies on disciplines such as epigraphy, the study of ancient inscribed texts, for evidence of the recorded past. However, these texts, "inscriptions", are often damaged over the centuries, and illegible parts of the text must be restored by specialists, known as epigraphists. This work presents Pythia, the first ancient text restoration model that recovers missing characters from a damaged text input using deep neural networks. Its architecture is carefully designed to handle long-term context information, and deal efficiently with missing or corrupted character and word representations. To train it, we wrote a non-trivial pipeline to convert PHI, the largest digital corpus of ancient Greek inscriptions, to machine actionable text, which we call PHI-ML. On PHI-ML, Pythia's predictions achieve a 30.1% character error rate, compared to the 57.3% of human epigraphists. Moreover, in 73.5% of cases the ground-truth sequence was among the Top-20 hypotheses of Pythia, which effectively demonstrates the impact of this assistive method on the field of digital epigraphy, and sets the state-of-the-art in ancient text restoration.