Qiujia Li

Knowledge Distillation for Neural Transducers from Large Self-Supervised Pre-trained Models

2021 · arXiv (Cornell University)

Self-supervised pre-training is an effective approach to leveraging a large amount of unlabelled data to reduce word error rates (WERs) of automatic speech recognition (ASR) systems. Since it is impractical to use large pre-trained models …

Modular Domain Adaptation for Conformer-Based Streaming ASR

2023 · arXiv (Cornell University)

Speech data from different domains has distinct acoustic and linguistic characteristics. It is common to train a single multidomain model such as a Conformer transducer for speech recognition on a mixture of data from all …