Multilingual Speech Translation with Efficient Finetuning of Pretrained Models
Authors:
- Xian LiXLXian Li
- Changhan WangCWChanghan Wang
- Yun TangYTYun Tang
- Chau TranCTChau Tran
- Yuqing TangYTYuqing Tang
- Juan PinoJPJuan Pino
- Alexei BaevskiABAlexei Baevski
- Alexis ConneauACAlexis Conneau
- Michael AuliMAMichael Auli
Preprint
Abstract
We present a simple yet effective approach to build multilingual speech-to-text (ST) translation by efficient transfer learning from pretrained speech encoder and text decoder. Our key finding is that a minimalistic LNA (LayerNorm and Attention) finetuning can achieve zero-shot crosslingual and cross-modality transfer ability by only finetuning less than 10% of the pretrained parameters. This enables effectively leveraging large pretrained models with low training cost. Using wav2vec 2.0 for acoustic modeling, and mBART for multilingual text generation, our approach advanced the new state-of-the-art for 34 translation directions (and surpassing cascaded ST for 23 of them) on large-scale multilingual ST benchmark CoVoST 2 (+6.4 BLEU on average across 15 En-X directions and +5.1 BLEU on average across 19 X-En directions). Our approach demonstrates strong zero-shot performance in a many-to-many multilingual model (+5.7 BLEU on average across 18 non-English directions), making it an appealing approach for attaining high-quality speech translation with improved parameter and data efficiency.Keywords
All Related Versions
This publication has 26 references indexed in Scilit:
- Multilingual Denoising Pre-training for Neural Machine TranslationTransactions of the Association for Computational Linguistics, 2020
- End-end Speech-to-Text Translation with Modality Agnostic Meta-Learning2020
- Unsupervised Pretraining Transfers Well Across Languages2020
- Bridging the Gap between Pre-Training and Fine-Tuning for End-to-End Speech TranslationProceedings of the AAAI Conference on Artificial Intelligence, 2020
- Fairseq S2T: Fast Speech-to-Text Modeling with Fairseq2020
- Curriculum Pre-training for End-to-End Speech Translation2020
- Improved Speech Representations with Multi-Target Autoregressive Predictive Coding2020
- A Comparative Study on End-to-End Speech to Text Translation2019
- Leveraging Weakly Supervised Data to Improve End-to-end Speech-to-text Translation2019
- End-to-End Automatic Speech Translation of Audiobooks2018