1 paper
Chen Xu, Xiaoqian Liu, Xiaowen Liu +9
Combining end-to-end speech translation (ST) and non-autoregressive (NAR) generation is promising in language and speech processing for their advantages of less error propagation a…