Improving low-resource CD-DNN-HMM using dropout and multilingual DNN training

By Yajie Miao, Florian Metze. 2013

72 citations recorded by the source corpus.

We investigate two strategies to improve the context-dependent deep neural network hidden Markov model (CD-DNN-HMM) in low-resource speech recognition. Although outperforming the conventional Gaussian mixture model (GMM) HMM on various tasks, CD-DNN-HMM acoustic modeling becomes challenging with limited transcribed speech, e.g., less than 10 hours. To resolve this issue, we firstly exploit dropout which prevents overfitting in DNN finetuning and improves model robustness under data sparseness. Then, the effectiveness of multilingual DNN training is evaluated when additional auxiliary languages are available. The hidden layer parameters of the target language are shared and learned over multiple languages. Experiments show that both strategies boost the recognition performance significantly. Combining them results in further reduction in word error rate, achieving 11.6% and 6.2% relative improvement on two limited data conditions.

Subfields: Artificial Intelligence, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Dong Yu
2. Li Deng
3. Florian Metze
4. Yajie Miao
5. Jinyu Li
6. Dongpeng Chen
7. Yoshua Bengio
8. Brian Kan-Wing Mak

In Signal Processing

1. Dong Yu
2. Li Deng
3. Florian Metze
4. Yajie Miao
5. Jinyu Li
6. Dongpeng Chen
7. Brian Kan-Wing Mak
8. Shourabh Rawat

Citations in the corpus

• Deep Learning: Methods and Applications (2014, methodology): “Subsequent work on applying dropout includes the study by Miao and Metze [243], where DNN-based speech recognition is constrained by low resources with sparse training data.”
• Deep maxout networks for low-resource speech recognition (2013, methodology): “Multilingual network training enables crosslanguage knowledge transfer and can happen either in pretraining [13] or in the fine-tuning stage [11, 12].”
• Multitask Learning of Deep Neural Networks for Low-Resource Speech Recognition (2015, methodology): “There are many other ways to do this, and perhaps the most well-known one is the dropout method [52] which had been applied successfully to low-resource ASR [53].”