Curriculum Pre-training for End-to-End Speech Translation
arXiv:2004.10093
Abstract
End-to-end speech translation poses a heavy burden on the encoder, because it has to transcribe, understand, and learn cross-lingual semantics simultaneously. To obtain a powerful encoder, traditional methods pre-train it on ASR data to capture speech features. However, we argue that pre-training the encoder only through simple speech recognition is not enough and high-level linguistic knowledge should be considered. Inspired by this, we propose a curriculum pre-training method that includes an elementary course for transcription learning and two advanced courses for understanding the utterance and mapping words in two languages. The difficulty of these courses is gradually increasing. Experiments show that our curriculum pre-training method leads to significant improvements on En-De and En-Fr speech translation benchmarks.
arXiv commentaccepted by ACL2020
References in corpus (4)
- SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition
- Listen and Translate: A Proof of Concept for End-to-End Speech-to-Text Translation
- Improving Transformer-based Speech Recognition Using Unsupervised Pre-training
- Semantic Mask for Transformer based End-to-End Speech Recognition
Cited by in corpus (5)
- Toward Transformer-Based Object Detection
- End-to-end Speech Translation via Cross-modal Progressive Training
- Adaptive Feature Selection for End-to-End Speech Translation
- Improving Reverberant Speech Separation with Multi-stage Training and Curriculum Learning
- A Technical Report: BUT Speech Translation Systems