# TDR: Tagalog Diacritic Restoration Dataset

arXiv: https://arxiv.org/abs/2610.17597 · HTML: https://mlsanity.com/paper/2610.17597/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 9 of its 50 references that are in the map.
- First posted on arXiv 13 Oct 2026.

## Builds on (most cited first, 9 of 9)

- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://mlsanity.com/paper/1907.11692/) (2019)
- [SEA-LION: Southeast Asian Languages in One Network](https://mlsanity.com/paper/2504.05747/) (2025)
- [Improving Large-scale Language Models and Resources for Filipino](https://mlsanity.com/paper/2111.06053/) (2021)
- [MOSS-TTS Technical Report](https://mlsanity.com/paper/2603.18090/) (2026)
- [Correcting diacritics and typos with a ByT5 transformer model](https://mlsanity.com/paper/2201.13242/) (2022)
- [calamanCy: A Tagalog Natural Language Processing Toolkit](https://mlsanity.com/paper/2311.07171/) (2023)
- [Restoring Arabic vowels through omission-tolerant dictionary lookup](https://mlsanity.com/paper/1905.04051/) (2019)
- [Towards Stress-Aware Sentence-Level Filipino G2P With Weakly-Supervised ByT5 Fine-Tuning](https://mlsanity.com/paper/2609.09974/) (2026)

## Abstract (arXiv)

Homographs are words that share identical spelling but differ in meaning and pronunciation. In Tagalog, these distinctions are disambiguated in writing through tuldik (diacritical marks), which are rarely preserved in contemporary Tagalog text. This scarcity of diacritics limits diacritic restoration, which is useful for downstream tasks such as prosody selection for text-to-speech synthesis. To address this gap, this study introduces TDR, an annotated corpus of 40 frequent Tagalog homographs constructed from news text. It also introduces a collocation-driven retrieval method to expand the dataset and mitigate severe class imbalance among minor variants. Finally, we show that TDR supplies a capability absent from pretrained models, and that this capability can be learned by small, inexpensive models as well as by fine-tuned LLMs.