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

Zhiyun Fan Linhao Dong Jun Zhang Lu Lu Zejun Ma

Published in: ICASSP (2024)

Keyphrases

automatic speech recognition
speech recognition
simulated annealing
training phase
training data
training set
training examples
training process
speaker recognition
data sets
supervised learning
online learning
test set
speech signal
speaker identification