Rethinking and Improving Multi-task Learning for End-to-end Speech Translation
This work addresses efficiency and performance in speech translation, which is incremental as it builds on existing multi-task learning approaches.
The paper tackled the problem of improving multi-task learning for end-to-end speech translation by investigating task consistency and proposing an improved method that bridges modal gaps in length and representation, achieving state-of-the-art results on the MuST-C dataset with a 20.8% reduction in training time compared to the current SOTA.
Significant improvements in end-to-end speech translation (ST) have been achieved through the application of multi-task learning. However, the extent to which auxiliary tasks are highly consistent with the ST task, and how much this approach truly helps, have not been thoroughly studied. In this paper, we investigate the consistency between different tasks, considering different times and modules. We find that the textual encoder primarily facilitates cross-modal conversion, but the presence of noise in speech impedes the consistency between text and speech representations. Furthermore, we propose an improved multi-task learning (IMTL) approach for the ST task, which bridges the modal gap by mitigating the difference in length and representation. We conduct experiments on the MuST-C dataset. The results demonstrate that our method attains state-of-the-art results. Moreover, when additional data is used, we achieve the new SOTA result on MuST-C English to Spanish task with 20.8% of the training time required by the current SOTA method.