Stoicheia: Character-Level Masked Diffusion for Ancient Greek Textual Restoration, Parsing, and Metrical Scansion
On the map: Diffusion models → Discrete and text diffusion
Abstract (arXiv)
We introduce Stoicheia, a 405M-parameter character-level masked-diffusion encoder for Ancient Greek whose input factors into five aligned, independently maskable planes: letters, word and sentence boundaries, diacritics, capitalization, and punctuation. A single backbone can therefore restore lacunae, re-segment, accentuate, and punctuate unspaced text without task-specific retokenization. We pretrain it on an open, revision-pinned corpus of 380M words and release eleven checkpoints: ten rotated, decontaminated folds, guaranteeing that for any given literary passage at least one released model has never seen its text, and one with no exposure to documentary texts. Three experiments - reconstruction of damaged inscriptions and papyri, morphosyntactic tagging and dependency parsing, and macronization with metrical scansion - each carry a matched random-initialization control, isolating what character-level diffusion pretraining contributes: 5.6 CER points on inscription reconstruction, 12.9 LAS on parsing, and 6.0 points of balanced accuracy on macronization. On Ithaca's own test split, with identical frozen samples and strict scoring, Stoicheia reduces character error relative to both prior state-of-the-art systems, from 24.6 (Ithaca) and 23.5 (its 2025 Aeneas-framework successor) to 15.5, and raises top-1 accuracy from 63.0 and 64.0 to 74.5.
Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Discrete and text diffusion, one of 14 directions in Diffusion models. Builds on 12 of its 31 references that are in the map. First posted on arXiv 7 Aug 2026.
Builds on
12 of 31 references.
Closest ancestors · 1
Same direction, newest first
Other references · 11
- 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
- 2019Unsupervised Cross-lingual Representation Learning at Scalecited by 3,907
- 2021Structured Denoising Diffusion Models in Discrete State-Spacescited by 1,665
- 2021ByT5: Towards a token-free future with pre-trained byte-to-byte modelscited by 504
- 2016Deep Biaffine Attention for Neural Dependency Parsingcited by 334
- 2021CANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representationcited by 229
- 2024Leak, Cheat, Repeat: Data Contamination and Evaluation Malpractices in Closed-Source LLMscited by 224
- 2019Restoring ancient text using deep learning: a case study on Greek epigraphycited by 25
- 2023Sentence Embedding Models for Ancient Greek Using Multilingual Knowledge Distillationcited by 6
- 2026Structure-Aware Text Recognition for Ancient Greek Critical Editionscited by 5
- 2024A State-of-the-Art Morphosyntactic Parser and Lemmatizer for Ancient Greekcited by 2
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2026When Simpler Is Better: Evaluating Translation Pipelines for Medieval Latin Manuscripts
- 2026Reading or Guessing? Visual Grounding Failures of Vision-Language Models for OCR in Ancient Greek Editions
- 2026Semantic Alignment across Ancient Egyptian Language Stages via Normalization-Aware Multitask Learning
- 2023Diff-Oracle: Deciphering Oracle Bone Scripts with Controllable Diffusion Model
- 2026FontFusion: Enhancing Generative Text in Diffusion Models with Typographic Conditioning
- 2026Elastic-dLLM: Position Preserving Context Compression and Augmentation of Diffusion LLMs
- 2026Intermediate Text Representation Guided Text-to-Image Generation for Enhancing One-and-Only Alignment
- 2026NAVIRA: Decoupled Stochastic Remasking for Masked Diffusion Language Models