SA-SOT: Speaker-Aware Serialized Output Training for Multi-Talker ASR.

Zhiyun Fan, Linhao Dong, Jun Zhang, Lu Lu, Zejun Ma

Browse the full ICASSP paper archive.