Towards Stress-Aware Sentence-Level Filipino G2P With Weakly-Supervised ByT5 Fine-Tuning

arXiv cs.CL Papers

Summary

This paper investigates weakly-supervised ByT5 fine-tuning for stress-aware sentence-level Filipino grapheme-to-phoneme conversion, achieving significant improvements in phoneme and character error rates on a manually-corrected test set.

arXiv:2609.09974v1 Announce Type: new Abstract: Grapheme-to-phoneme conversion (G2P) refers to the task of converting a sequence of graphemes to a corresponding sequence of phonemes. While Filipino G2P is fairly straightforward due to its shallow orthography, the inclusion of prosodic features such as stress adds a layer of complexity that requires sentence-level context instead of single-word inputs. However, sentence-level data for Filipino typically do not include phoneme transcriptions, posing a challenge for training G2P models. As such, we investigate how to obtain sentence-level phoneme data for Filipino using available data and compare the resulting models with multilingual word-level G2P as well as measure how accurately they predict stress marker position for Filipino. We propose fine-tuning a ByT5-based model, pre-trained on multilingual word-level G2P data, on three sentence-level G2P datasets annotated with an LLM-assisted pipeline guided by data from Wiktionary. This approach produces models that perform well on the G2P task, achieving at best around 0.54% PER and 2.50% CER, a significant decrease compared to base model PER at around 19.74%, on a manually-corrected test set. The model is able to correctly classify most of the main stress classes in Filipino, but struggles particularly with malumi words. We show that a ByT5-based model performs well at sentence-level Filipino G2P and offers strong potential for Filipino homograph disambiguation.
Original Article
View Cached Full Text

Cached at: 09/10/26, 08:17 AM

# Towards Stress-Aware Sentence-Level Filipino G2P With Weakly-Supervised ByT5 Fine-Tuning
Source: [https://arxiv.org/abs/2609.09974](https://arxiv.org/abs/2609.09974)
[View PDF](https://arxiv.org/pdf/2609.09974)

> Abstract:Grapheme\-to\-phoneme conversion \(G2P\) refers to the task of converting a sequence of graphemes to a corresponding sequence of phonemes\. While Filipino G2P is fairly straightforward due to its shallow orthography, the inclusion of prosodic features such as stress adds a layer of complexity that requires sentence\-level context instead of single\-word inputs\. However, sentence\-level data for Filipino typically do not include phoneme transcriptions, posing a challenge for training G2P models\. As such, we investigate how to obtain sentence\-level phoneme data for Filipino using available data and compare the resulting models with multilingual word\-level G2P as well as measure how accurately they predict stress marker position for Filipino\. We propose fine\-tuning a ByT5\-based model, pre\-trained on multilingual word\-level G2P data, on three sentence\-level G2P datasets annotated with an LLM\-assisted pipeline guided by data from Wiktionary\. This approach produces models that perform well on the G2P task, achieving at best around 0\.54% PER and 2\.50% CER, a significant decrease compared to base model PER at around 19\.74%, on a manually\-corrected test set\. The model is able to correctly classify most of the main stress classes in Filipino, but struggles particularly with malumi words\. We show that a ByT5\-based model performs well at sentence\-level Filipino G2P and offers strong potential for Filipino homograph disambiguation\.

## Submission history

From: Lorenz Bernard Marqueses \[[view email](https://arxiv.org/show-email/8ce27d06/2609.09974)\] **\[v1\]**Wed, 9 Sep 2026 10:02:25 UTC \(1,116 KB\)

Similar Articles

Graph-Based Phonetic Error Correction of Noisy ASR

arXiv cs.CL

Proposes G-SPIN, a lightweight framework that combines phonetic graph modeling with contextual language understanding for correcting ASR errors, using a GNN to generate phonetically plausible candidate tokens, an MLM for local scoring, and an LLM for final re-ranking, all operating at inference time.