# Contextualized Translation of Automatically Segmented Speech

arXiv: https://arxiv.org/abs/2008.02270 · HTML: https://mlsanity.com/paper/2008.02270/

## Key facts

- Cited by 12 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 5 Aug 2020.

## Built on by (most cited first, 12 of 12)

- [StreamAtt: Direct Streaming Speech-to-Text Translation with Attention-based Audio History Selection](https://mlsanity.com/paper/2406.06097/) (2024)
- [How "Real" is Your Real-Time Simultaneous Speech-to-Text Translation System?](https://mlsanity.com/paper/2412.18495/) (2024)
- [Efficient yet Competitive Speech Translation: FBK@IWSLT2022](https://mlsanity.com/paper/2205.02629/) (2022)
- [Non-autoregressive End-to-end Speech Translation with Parallel Autoregressive Rescoring](https://mlsanity.com/paper/2109.04411/) (2021)
- [Fast-MD: Fast Multi-Decoder End-to-End Speech Translation with Non-Autoregressive Hidden Intermediates](https://mlsanity.com/paper/2109.12804/) (2021)
- [Don't Discard Fixed-Window Audio Segmentation in Speech-to-Text Translation](https://mlsanity.com/paper/2210.13363/) (2022)
- [Dealing with training and test segmentation mismatch: FBK@IWSLT2021](https://mlsanity.com/paper/2106.12607/) (2021)
- [SegAugment: Maximizing the Utility of Speech Translation Data with Segmentation-based Augmentations](https://mlsanity.com/paper/2212.09699/) (2022)
- [Dodging the Data Bottleneck: Automatic Subtitling with Automatically Segmented ST Corpora](https://mlsanity.com/paper/2209.10608/) (2022)
- [Named Entity Detection and Injection for Direct Speech Translation](https://mlsanity.com/paper/2210.11981/) (2022)
- [Long-form Simultaneous Speech Translation: Thesis Proposal](https://mlsanity.com/paper/2310.11141/) (2023)
- [ESPnet-ST IWSLT 2021 Offline Speech Translation System](https://mlsanity.com/paper/2107.00636/) (2021)

## Abstract (arXiv)

Direct speech-to-text translation (ST) models are usually trained on corpora segmented at sentence level, but at inference time they are commonly fed with audio split by a voice activity detector (VAD). Since VAD segmentation is not syntax-informed, the resulting segments do not necessarily correspond to well-formed sentences uttered by the speaker but, most likely, to fragments of one or more sentences. This segmentation mismatch degrades considerably the quality of ST models' output. So far, researchers have focused on improving audio segmentation towards producing sentence-like splits. In this paper, instead, we address the issue in the model, making it more robust to a different, potentially sub-optimal segmentation. To this aim, we train our models on randomly segmented data and compare two approaches: fine-tuning and adding the previous segment as context. We show that our context-aware solution is more robust to VAD-segmented input, outperforming a strong base model and the fine-tuning on different VAD segmentations of an English-German test set by up to 4.25 BLEU points.