Label-Synchronous Neural Transducer for E2E Simultaneous Speech Translation
This addresses the challenge of streaming and re-ordering for end-to-end simultaneous speech translation, offering incremental improvements in quality and latency for applications like real-time translation.
The paper tackles the problem of simultaneous speech translation by proposing the LS-Transducer-SST, which dynamically emits translation tokens and controls latency, resulting in a better quality-latency trade-off, such as a 3.1/2.9 point BLEU increase relative to CAAT at similar latency and a 1.4 s reduction in average lagging latency relative to Wait-k.
While the neural transducer is popular for online speech recognition, simultaneous speech translation (SST) requires both streaming and re-ordering capabilities. This paper presents the LS-Transducer-SST, a label-synchronous neural transducer for SST, which naturally possesses these two properties. The LS-Transducer-SST dynamically decides when to emit translation tokens based on an Auto-regressive Integrate-and-Fire (AIF) mechanism. A latency-controllable AIF is also proposed, which can control the quality-latency trade-off either only during decoding, or it can be used in both decoding and training. The LS-Transducer-SST can naturally utilise monolingual text-only data via its prediction network which helps alleviate the key issue of data sparsity for E2E SST. During decoding, a chunk-based incremental joint decoding technique is designed to refine and expand the search space. Experiments on the Fisher-CallHome Spanish (Es-En) and MuST-C En-De data show that the LS-Transducer-SST gives a better quality-latency trade-off than existing popular methods. For example, the LS-Transducer-SST gives a 3.1/2.9 point BLEU increase (Es-En/En-De) relative to CAAT at a similar latency and a 1.4 s reduction in average lagging latency with similar BLEU scores relative to Wait-k.