# Mitigating Structural Noise in Low-Resource S2TT: An Optimized Cascaded Nepali-English Pipeline with Punctuation Restoration

arXiv: https://arxiv.org/abs/2602.21647 · HTML: https://mlsanity.com/paper/2602.21647/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 11 of its 23 references that are in the map.
- First posted on arXiv 25 Feb 2026.

## Builds on (most cited first, 11 of 11)

- [Multilingual Denoising Pre-training for Neural Machine Translation](https://mlsanity.com/paper/2001.08210/) (2020)
- [Six Challenges for Neural Machine Translation](https://mlsanity.com/paper/1706.03872/) (2017)
- [Transfer Learning for Low-Resource Neural Machine Translation](https://mlsanity.com/paper/1604.02201/) (2016)
- [Survey of Low-Resource Machine Translation](https://mlsanity.com/paper/2109.00486/) (2021)
- [Cascade versus Direct Speech Translation: Do the Differences Still Make a Difference?](https://mlsanity.com/paper/2106.01045/) (2021)
- [Applying Wav2vec2.0 to Speech Recognition in Various Low-resource Languages](https://mlsanity.com/paper/2012.12121/) (2020)
- [Wav2vec-S: Semi-Supervised Pre-Training for Low-Resource ASR](https://mlsanity.com/paper/2110.04484/) (2021)
- [Domain-adaptative Continual Learning for Low-resource Tasks: Evaluation on Nepali](https://mlsanity.com/paper/2412.13860/) (2024)
- [Direct Speech to Speech Translation: A Review](https://mlsanity.com/paper/2503.04799/) (2025)
- [Automatic speech recognition for the Nepali language using CNN, bidirectional LSTM and ResNet](https://mlsanity.com/paper/2406.17825/) (2024)
- [Pivot Language for Low-Resource Machine Translation](https://mlsanity.com/paper/2505.14553/) (2025)

## Abstract (arXiv)

Cascaded speech-to-text translation (S2TT) systems for low-resource languages can suffer from structural noise, particularly the loss of punctuation during the Automatic Speech Recognition (ASR) phase. This research investigates the impact of such noise on Nepali-to-English translation and proposes an optimized pipeline to mitigate quality degradation. We first establish highly proficient ASR and NMT components: a Wav2Vec2-XLS-R-300m model achieved a state-of-the-art 2.72% CER on OpenSLR-54, and a multi-stage fine-tuned MarianMT model reached a 28.32 BLEU score on the FLORES-200 benchmark. We empirically investigate the influence of punctuation loss, demonstrating that unpunctuated ASR output significantly degrades translation quality, causing a massive 20.7% relative BLEU drop on the FLORES benchmark. To overcome this, we propose and evaluate an intermediate Punctuation Restoration Module (PRM). The final S2TT pipeline was tested across three configurations on a custom dataset. The optimal configuration, which applied the PRM directly to ASR output, achieved a 4.90 BLEU point gain over the direct ASR-to-NMT baseline (BLEU 36.38 vs. 31.48). This improvement was validated by human assessment, which confirmed the optimized pipeline's superior Adequacy (3.673) and Fluency (3.804) with inter-rater reliability (Krippendorff's α ≥ 0.723). This work validates that targeted punctuation restoration is the most effective intervention for mitigating structural noise in the Nepali S2TT pipeline. It establishes an optimized baseline and demonstrates a critical architectural insight for developing cascaded speech translation systems for similar low-resource languages.