mlsanityData to 14 Oct 2026

Leveraging Audio-LLMs to Filter Speech-to-Speech Training Data

arXiv 2606.13507 ↗First posted 11 Jun 2026cs.CLcited by 0

Not on the map. Why?

Abstract (arXiv)

Large-scale mined corpora provide abundant training data for end-to-end speech-to-speech translation (S2ST) but may contain noise, misalignment, and semantic errors. Filtering noisy data is crucial to maintain robust speech translation performance. We study how to train an audio-language model to make keep/drop decisions on paired speech directly from audio. To obtain reliable supervision without manual labels, we adopt a scalable two-stage Rank-to-Distill strategy. A lightweight ranker generates keep/drop pseudo-labels from noisy speech pairs, then trains an audio large language model to predict keep/drop directly from raw paired speech. The resulting model jointly captures acoustic fidelity and cross-lingual semantic consistency for the selection of speech-conditioned data. Experiments on CVSS-C and SpeechMatrix show consistent improvements over unfiltered training, yielding up to +1.4 ASR-BLEU for end-to-end S2ST.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 23 of its 28 references that are in the map. First posted on arXiv 11 Jun 2026.

Builds on

23 of 28 references.

References · 23

  1. 2025Qwen3 Technical Reportcited by 8,820
  2. 2022Robust Speech Recognition via Large-Scale Weak Supervisioncited by 4,057
  3. 2019Self-training with Noisy Student improves ImageNet classificationcited by 1,311
  4. 2020BLEURT: Learning Robust Metrics for Text Generationcited by 1,203
  5. 2024Qwen2-Audio Technical Reportcited by 560
  6. 2023Large Language Models Are State-of-the-Art Evaluators of Translation Qualitycited by 420
  7. 2022UTMOS: UTokyo-SaruLab System for VoiceMOS Challenge 2022cited by 390
  8. 2024Audio Flamingo: A Novel Audio Language Model with Few-Shot Learning and Dialogue Abilitiescited by 186
  9. 2021Direct speech-to-speech translation with discrete unitscited by 149
  10. 2017Sequence-to-Sequence Models Can Directly Translate Foreign Speechcited by 121
  11. 2019Direct speech-to-speech translation with a sequence-to-sequence modelcited by 118
  12. 2022CVSS Corpus and Massively Multilingual Speech-to-Speech Translationcited by 71
11 more
  1. 2019Curriculum Learning for Domain Adaptation in Neural Machine Translationcited by 59
  2. 2024LLaMAX: Scaling Linguistic Horizons of LLM by Enhancing Translation Capabilities Beyond 100 Languagescited by 48
  3. 2018Denoising Neural Machine Translation Training with Trusted Data and Online Data Selectioncited by 43
  4. 2018Effective Parallel Corpus Mining using Bilingual Sentence Embeddingscited by 39
  5. 2022SpeechMatrix: A Large-Scale Mined Corpus of Multilingual Speech-to-Speech Translationscited by 35
  6. 2022TranSpeech: Speech-to-Speech Translation With Bilateral Perturbationcited by 34
  7. 2019Low-Resource Corpus Filtering using Multilingual Sentence Embeddingscited by 25
  8. 2022Pseudo Label Is Better Than Human Labelcited by 25
  9. 2025Audio Large Language Models Can Be Descriptive Speech Quality Evaluatorscited by 25
  10. 2022Brouhaha: multi-task training for voice activity detection, speech-to-noise ratio, and C50 room acoustics estimationcited by 23
  11. 2024A Case Study on Filtering for End-to-End Speech Translationcited by 2

Built on by

Not cited yet.