Low-Resource Language Discrimination toward Chinese Dialects with Transfer Learning and Data Augmentation
Authors:
Open Access
- 31 October 2021
- journal article
- research article
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Asian and Low-Resource Language Information Processing
- Vol. 21 (2) , 1-21
Abstract
Chinese dialects discrimination is a challenging natural language processing task due to scarce annotation resource. In this article, we develop a novel Chinese dialects discrimination framework with transfer learning and data augmentation (CDDTLDA) in order to overcome the shortage of resources. To be more specific, we first use a relatively larger Chinese dialects corpus to train a source-side automatic speech recognition (ASR) model. Then, we adopt a simple but effective data augmentation method (i.e., speed, pitch, and noise disturbance) to augment the target-side low-resource Chinese dialects, and fine-tune another target ASR model based on the previous source-side ASR model. Meanwhile, the potential common semantic features between source-side and target-side ASR models can be captured by using self-attention mechanism. Finally, we extract the hidden semantic representation in the target ASR model to conduct Chinese dialects discrimination. Our extensive experimental results demonstrate that our model significantly outperforms state-of-the-art methods on two benchmark Chinese dialects corpora.Keywords
Funding Information
- National Natural Science Foundation of China (61772246, 62162031, 62066020, and 61876074)
- Natural Science Foundation of Jiangxi Province (20192ACBL21030 and 20192AE191005)
This publication has 19 references indexed in Scilit:
- Spoken Language Recognition using X-vectors2018
- Findings of the VarDial Evaluation Campaign 20172017
- Discriminating between Similar Languages Using a Combination of Typed and Untyped Character N-grams and Words2017
- The NRC System for Discriminating Similar Languages2014
- Domain Adaptation via Transfer Component AnalysisIEEE Transactions on Neural Networks, 2010
- A Survey on Transfer LearningIEEE Transactions on Knowledge and Data Engineering, 2009
- Boosting for transfer learning2007
- Language identification from small text samples*Journal of Quantitative Linguistics, 2006
- Testing the correlation of word error rate and perplexitySpeech Communication, 2002
- Comparison of parametric representations for monosyllabic word recognition in continuously spoken sentencesIEEE Transactions on Acoustics, Speech, and Signal Processing, 1980