# ScribeTokens: Fixed-Vocabulary Tokenization of Digital Ink

arXiv: https://arxiv.org/abs/2603.02805 · HTML: https://mlsanity.com/paper/2603.02805/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 25 of its 36 references that are in the map.
- First posted on arXiv 3 Mar 2026.

## Builds on (most cited first, 25 of 25)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://mlsanity.com/paper/1910.10683/) (2019)
- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://mlsanity.com/paper/2104.09864/) (2021)
- [Neural Machine Translation of Rare Words with Subword Units](https://mlsanity.com/paper/1508.07909/) (2015)
- [GLU Variants Improve Transformer](https://mlsanity.com/paper/2002.05202/) (2020)
- [Root Mean Square Layer Normalization](https://mlsanity.com/paper/1910.07467/) (2019)
- [Quantifying Attention Flow in Transformers](https://mlsanity.com/paper/2005.00928/) (2020)
- [Generating Sequences With Recurrent Neural Networks](https://mlsanity.com/paper/1308.0850/) (2013)
- [Jukebox: A Generative Model for Music](https://mlsanity.com/paper/2005.00341/) (2020)
- [CosyVoice: A Scalable Multilingual Zero-shot Text-to-speech Synthesizer based on Supervised Semantic Tokens](https://mlsanity.com/paper/2407.05407/) (2024)
- [A Neural Representation of Sketch Drawings](https://mlsanity.com/paper/1704.03477/) (2017)
- [Multimodal Trajectory Predictions for Autonomous Driving using Deep Convolutional Networks](https://mlsanity.com/paper/1809.10732/) (2018)
- [Overcoming Limitations of Mixture Density Networks: A Sampling and Fitting Framework for Multimodal Future Prediction](https://mlsanity.com/paper/1906.03631/) (2019)
- [Sketchformer: Transformer-based Representation for Sketched Structure](https://mlsanity.com/paper/2002.10381/) (2020)
- [Sketch-BERT: Learning Sketch Bidirectional Encoder Representation from Transformers by Self-supervised Learning of Sketch Gestalt](https://mlsanity.com/paper/2005.09159/) (2020)
- [DeepWriting: Making Digital Ink Editable via Deep Generative Modeling](https://mlsanity.com/paper/1801.08379/) (2018)
- [Vectorization and Rasterization: Self-Supervised Learning for Sketch and Handwriting](https://mlsanity.com/paper/2103.13716/) (2021)
- [Disentangling Writer and Character Styles for Handwriting Generation](https://mlsanity.com/paper/2303.14736/) (2023)
- [Drawing and Recognizing Chinese Characters with Recurrent Neural Network](https://mlsanity.com/paper/1606.06539/) (2016)
- [Fast Multi-language LSTM-based Online Handwriting Recognition](https://mlsanity.com/paper/1902.10525/) (2019)
- [Representing Online Handwriting for Recognition in Large Vision-Language Models](https://mlsanity.com/paper/2402.15307/) (2024)
- [SketchGPT: Autoregressive Modeling for Sketch Generation and Recognition](https://mlsanity.com/paper/2405.03099/) (2024)

## Abstract (arXiv)

Digital ink -- the coordinate stream captured from stylus or touch input -- lacks a unified representation. Continuous vector representations produce long sequences and suffer from training instability, while existing token representations require large vocabularies, face out-of-vocabulary issues, and underperform vectors on recognition. We propose ScribeTokens, a tokenization that decomposes pen movement into unit pixel steps. Together with two pen-state tokens, this fixed 10-token base vocabulary suffices to represent any digital ink and enables aggressive BPE compression. On handwritten text generation, ScribeTokens dramatically outperforms vectors (17.33% vs. 70.29% CER), showing tokens are far more effective for generation. On recognition, ScribeTokens is the only token representation to outperform vectors without pretraining. We further introduce next-ink-token prediction as a self-supervised pretraining strategy, which consistently improves recognition across all token-based models and accelerates convergence by up to 83x. With pretraining, ScribeTokens achieves the best recognition results across all representations on both datasets (8.27% CER on IAM, 9.83% on DeepWriting).