Lightweight Adapter Tuning for Multilingual Speech Translation
Preprint
Abstract
Adapter modules were recently introduced as an efficient alternative to fine-tuning in NLP. Adapter tuning consists in freezing pretrained parameters of a model and injecting lightweight modules between layers, resulting in the addition of only a small number of task-specific trainable parameters. While adapter tuning was investigated for multilingual neural machine translation, this paper proposes a comprehensive analysis of adapters for multilingual speech translation (ST). Starting from different pre-trained models (a multilingual ST trained on parallel data or a multilingual BART (mBART) trained on non-parallel multilingual data), we show that adapters can be used to: (a) efficiently specialize ST to specific language pairs with a low extra cost in terms of parameters, and (b) transfer from an automatic speech recognition (ASR) task and an mBART pre-trained model to a multilingual ST task. Experiments show that adapter tuning offer competitive results to full fine-tuning, while being much more parameter-efficient.Keywords
All Related Versions
This publication has 15 references indexed in Scilit:
- Adaptable Multi-Domain Language Model for Transformer ASR2021
- AdapterFusion: Non-Destructive Task Composition for Transfer Learning2021
- Recipes for Adapting Pre-trained Monolingual and Multilingual Models to Machine Translation2021
- Multilingual Speech Translation from Efficient Finetuning of Pretrained Models2021
- End-to-End Speech Translation with Pre-trained Models and Adapters: UPC at IWSLT 20212021
- Multilingual Denoising Pre-training for Neural Machine TranslationTransactions of the Association for Computational Linguistics, 2020
- Monolingual Adapters for Zero-Shot Neural Machine Translation2020
- BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension2020
- ESPnet-ST: All-in-One Speech Translation Toolkit2020
- Backpropagation Applied to Handwritten Zip Code RecognitionNeural Computation, 1989