Very deep convolutional neural networks for robust speech recognition

By Yanmin Qian, Philip C. Woodland. 2016

83 citations recorded by the source corpus.

This paper describes the extension and optimisation of our previous work on very deep convolutional neural networks (CNNs) for effective recognition of noisy speech in the Aurora 4 task. The appropriate number of convolutional layers, the sizes of the filters, pooling operations and input feature maps are all modified: the filter and pooling sizes are reduced and dimensions of input feature maps are extended to allow adding more convolutional layers. Furthermore appropriate input padding and input feature map selection strategies are developed. In addition, an adaptation framework using joint training of very deep CNN with auxiliary features i-vector and fMLLR features is developed. These modifications give substantial word error rate reductions over the standard CNN used as baseline. Finally the very deep CNN is combined with an LSTM-RNN acoustic model and it is shown that state-level weighted log likelihood score combination in a joint acoustic model decoding scheme is very effective. On the Aurora 4 task, the very deep CNN achieves a WER of 8.81%, further 7.99% with auxiliary feature joint training, and 7.09% with LSTM-RNN joint decoding.

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. Yanmin Qian
2. Xuankai Chang
3. Dong Yu
4. Dan Su
5. Zhao You
6. Shulin Feng
7. Li Deng
8. Jinyu Li

In Signal Processing

1. Yanmin Qian
2. Xuankai Chang
3. Dong Yu
4. Dan Su
5. Zhao You
6. Shulin Feng
7. Li Deng
8. Jinyu Li

Citations in the corpus

• Single-Channel Multi-talker Speech Recognition with Permutation Invariant Training (2017, background): “Thanks to the significant progresses made in the recent years [1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], the ASR systems now surpassed the threshold for adoption in many real-world scenarios and enabled services such as Microsoft…”
• SpeechMoE: Scaling to Large Acoustic Models with Dynamic Routing Mixture of Experts (2021, methodology): “Various types of neural network architectures have been employed in ASR systems, such as convolutional neural networks (CNNs) [3, 4], long short-term memory (LSTM) [5], gated recurrent unit[6], time-delayed neural network [7], feedforward sequential memory networks (FSMN) [8], etc.”