Low-resource Language Discrimination Towards Chinese Dialects with Transfer learning and Data Augmentation

The article presents a novel framework for Chinese dialect discrimination, termed CDDTLDA, which utilizes transfer learning and data augmentation techniques to address the challenges posed by limited annotated resources. The approach involves training a source ASR model on a larger corpus and employing data augmentation methods such as speed, pitch, and noise disturbances to enhance the target ASR model. Experimental results indicate that this method significantly surpasses existing state-of-the-art models on benchmark datasets, highlighting its potential impact on improving NLP tasks in low-resource settings.