1 paper
Haozhe Shan, Albert Gu, Zhong Meng +3
Online speech recognition, where the model only accesses context to the left, is an important and challenging use case for ASR systems. In this work, we investigate augmenting neur…