# Cascaded Models With Cyclic Feedback For Direct Speech Translation

arXiv: https://arxiv.org/abs/2010.11153 · HTML: https://mlsanity.com/paper/2010.11153/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 21 Oct 2020.

## Built on by (most cited first, 9 of 9)

- [STEMM: Self-learning with Speech-text Manifold Mixup for Speech Translation](https://mlsanity.com/paper/2203.10426/) (2022)
- [Sample, Translate, Recombine: Leveraging Audio Alignments for Data Augmentation in End-to-end Speech Translation](https://mlsanity.com/paper/2203.08757/) (2022)
- [Non-Parametric Domain Adaptation for End-to-End Speech Translation](https://mlsanity.com/paper/2205.11211/) (2022)
- [LLaST: Improved End-to-end Speech Translation System Leveraged by Large Language Models](https://mlsanity.com/paper/2407.15415/) (2024)
- [Align, Write, Re-order: Explainable End-to-End Speech Translation via Operation Sequence Generation](https://mlsanity.com/paper/2211.05967/) (2022)
- [Improving speech translation by fusing speech and text](https://mlsanity.com/paper/2305.14042/) (2023)
- [Speech Translation Refinement using Large Language Models](https://mlsanity.com/paper/2501.15090/) (2025)
- [R-BI: Regularized Batched Inputs enhance Incremental Decoding Framework for Low-Latency Simultaneous Speech Translation](https://mlsanity.com/paper/2401.05700/) (2024)
- [Interpretable Style Takagi-Sugeno-Kang Fuzzy Clustering](https://mlsanity.com/paper/2504.05125/) (2025)

## Abstract (arXiv)

Direct speech translation describes a scenario where only speech inputs and corresponding translations are available. Such data are notoriously limited. We present a technique that allows cascades of automatic speech recognition (ASR) and machine translation (MT) to exploit in-domain direct speech translation data in addition to out-of-domain MT and ASR data. After pre-training MT and ASR, we use a feedback cycle where the downstream performance of the MT system is used as a signal to improve the ASR system by self-training, and the MT component is fine-tuned on multiple ASR outputs, making it more tolerant towards spelling variations. A comparison to end-to-end speech translation using components of identical architecture and the same data shows gains of up to 3.8 BLEU points on LibriVoxDeEn and up to 5.1 BLEU points on CoVoST for German-to-English speech translation.