The USTC-NELSLIP Systems for Simultaneous Speech Translation Task at IWSLT 2021
This work addresses the problem of real-time translation for speech and text applications, showing incremental improvements over existing approaches.
The paper tackled simultaneous speech translation by proposing the Cross Attention Augmented Transducer (CAAT) model, which achieved better quality-latency trade-offs than previous methods, with S2T and T2T systems improving by an average of 11.3 and 4.6 BLEU, respectively, compared to last year's systems.
This paper describes USTC-NELSLIP's submissions to the IWSLT2021 Simultaneous Speech Translation task. We proposed a novel simultaneous translation model, Cross Attention Augmented Transducer (CAAT), which extends conventional RNN-T to sequence-to-sequence tasks without monotonic constraints, e.g., simultaneous translation. Experiments on speech-to-text (S2T) and text-to-text (T2T) simultaneous translation tasks shows CAAT achieves better quality-latency trade-offs compared to \textit{wait-k}, one of the previous state-of-the-art approaches. Based on CAAT architecture and data augmentation, we build S2T and T2T simultaneous translation systems in this evaluation campaign. Compared to last year's optimal systems, our S2T simultaneous translation system improves by an average of 11.3 BLEU for all latency regimes, and our T2T simultaneous translation system improves by an average of 4.6 BLEU.