UniSpeech: Unified Speech Representation Learning with Labeled and Unlabeled Data

Abstract

In this paper, we propose a unified pre-training approach called UniSpeech to learn speech representations with both unlabeled and labeled data, in which supervised phonetic CTC learning and phonetically-aware contrastive self-supervised learning are conducted in a multi-task learning manner. The resultant representations can capture information more correlated with phonetic structures and improve the generalization across languages and domains. We evaluate the effectiveness of UniSpeech for cross-lingual representation learning on public CommonVoice corpus. The results show that UniSpeech outperforms self-supervised pretraining and supervised transfer learning for speech recognition by a maximum of 13.4% and 17.8% relative phone error rate reductions respectively (averaged over all testing languages). The transferability of UniSpeech is also demonstrated on a domain-shift speech recognition task, i.e., a relative word error rate reduction of 6% against the previous approach.

accepted by ICML2021

References in corpus (13)

• Categorical Reparameterization with Gumbel-Softmax
• Sequence Transduction with Recurrent Neural Networks
• Achieving Human Parity in Conversational Speech Recognition
• Mockingjay: Unsupervised Speech Representation Learning with Deep Bidirectional Transformer Encoders
• vq-wav2vec: Self-Supervised Learning of Discrete Speech Representations
• TED-LIUM 3: twice as much data and corpus repartition for experiments on speaker adaptation
• Common Voice: A Massively-Multilingual Speech Corpus
• fairseq: A Fast, Extensible Toolkit for Sequence Modeling
• Deep Contextualized Acoustic Representations For Semi-Supervised Speech Recognition
• ESPnet: End-to-End Speech Processing Toolkit
• BERTphone: Phonetically-Aware Encoder Representations for Utterance-Level Speaker and Language Recognition
• On the Comparison of Popular End-to-End Models for Large Scale Speech Recognition
• Transformer in action: a comparative study of transformer-based acoustic models for large scale speech recognition applications

Cited by in corpus (3)

• UniSpeech at scale: An Empirical Study of Pre-training Method on Large-Scale Speech Recognition Dataset
• XLST: Cross-lingual Self-training to Learn Multilingual Representation for Low Resource Speech Recognition
• End-to-end Speech Translation via Cross-modal Progressive Training