JSTSP Volume 16 Issue 6

Recently, self-supervised learning (SSL) from unlabelled speech data has gained increased attention in the automatic speech recognition (ASR) community. Typical SSL methods include autoregressive predictive coding (APC), Wav2vec2.0, and hidden unit BERT (HuBERT). However, SSL models are biased to the pretraining data. When SSL models are finetuned with data from another domain, domain shifting occurs and might cause limited knowledge transfer for downstream tasks.

Speech self-supervised learning has attracted much attention due to its promising performance in multiple downstream tasks, and has become a new growth engine for speech recognition in low-resource languages. In this paper, we exploit and analyze a series of wav2vec pre-trained models for speech recognition in 15 low-resource languages in the OpenASR21 Challenge.

Although supervised deep learning has revolutionized speech and audio processing, it has necessitated the building of specialist models for individual tasks and application scenarios. It is likewise difficult to apply this to dialects and languages for which only limited labeled data is available. Self-supervised representation learning methods promise a single universal model that would benefit a wide variety of tasks and domains. 

Although supervised deep learning has revolutionized speech and audio processing, it has necessitated the building of specialist models for individual tasks and application scenarios. It is likewise difficult to apply this to dialects and languages for which only limited labeled data is available. Self-supervised representation learning methods promise a single universal model that would benefit a wide variety of tasks and domains. 

The papers in this special section focus on self-supervised learning for speech and audio processing. A current trend in the machine learning community is the adoption of self-supervised approaches to pretrain deep networks. Self-supervised learning utilizes proxy-supervised learning tasks (or pretext tasks) - for example, distinguishing parts of the input signal from distractors or reconstructing masked input segments conditioned on unmasked segments—to obtain training data from unlabeled corpora.