End-to-End Speech-Translation with Knowledge Distillation: FBK@IWSLT2020
This work addresses speech translation for TED talks, showing incremental improvements in BLEU scores with a focus on handling segmented versus raw audio data.
The paper tackled the IWSLT 2020 offline speech translation task by developing an end-to-end Transformer-based model for translating English TED talks audio to German text, achieving 29 BLEU on the MuST-C En-De test set and 23.7 BLEU on VAD-segmented data.
This paper describes FBK's participation in the IWSLT 2020 offline speech translation (ST) task. The task evaluates systems' ability to translate English TED talks audio into German texts. The test talks are provided in two versions: one contains the data already segmented with automatic tools and the other is the raw data without any segmentation. Participants can decide whether to work on custom segmentation or not. We used the provided segmentation. Our system is an end-to-end model based on an adaptation of the Transformer for speech data. Its training process is the main focus of this paper and it is based on: i) transfer learning (ASR pretraining and knowledge distillation), ii) data augmentation (SpecAugment, time stretch and synthetic data), iii) combining synthetic and real data marked as different domains, and iv) multi-task learning using the CTC loss. Finally, after the training with word-level knowledge distillation is complete, our ST models are fine-tuned using label smoothed cross entropy. Our best model scored 29 BLEU on the MuST-C En-De test set, which is an excellent result compared to recent papers, and 23.7 BLEU on the same data segmented with VAD, showing the need for researching solutions addressing this specific data condition.