Dual-Path Transformer Network: Direct Context-Aware Modeling for End-to-End Monaural Speech Separation

By Jing-jing Chen, Qi-rong Mao, Dong Liu. Interspeech, 2020

348 citations recorded by the source corpus.

The dominant speech separation models are based on complex recurrent or convolution neural network that model speech sequences indirectly conditioning on context, such as passing information through many intermediate states in recurrent neural network, leading to suboptimal separation performance. In this paper, we propose a dual-path transformer network (DPTNet) for end-to-end speech separation, which introduces direct context-awareness in the modeling for speech sequences. By introduces a improved transformer, elements in speech sequences can interact directly, which enables DPTNet can model for the speech sequences with direct context-awareness. The improved transformer in our approach learns the order information of the speech sequences without positional encodings by incorporating a recurrent neural network into the original transformer. In addition, the structure of dual paths makes our model efficient for extremely long speech sequence modeling. Extensive experiments on benchmark datasets show that our approach outperforms the current state-of-the-arts (20.6 dB SDR on the public WSj0-2mix data corpus).

Subfields: Artificial Intelligence, Signal Processing

View Paper: https://www.semanticscholar.org/paper/e2e055f18b01eac6eb14f7462c3708ad9f073316

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. Jinyu Li
2. Zhuo Chen
3. Shujie Liu
4. Takuya Yoshioka
5. Yu Wu
6. Sanyuan Chen
7. Jian Wu
8. Mirco Ravanelli

In Signal Processing

1. Jinyu Li
2. Zhuo Chen
3. Takuya Yoshioka
4. Shujie Liu
5. Yu Wu
6. Sanyuan Chen
7. Mirco Ravanelli
8. Jian Wu

Citations in the corpus

• Attention Is All You Need In Speech Separation (2020, positive): “An attempt to integrate transformers into the speech separation pipeline has been recently done in [22] where the proposed Dual-Path Transformer Network (DPTNet) is shown to outperform the standard DPRNN.”
• Ultra Fast Speech Separation Model with Teacher Student Learning (2021, positive): “Recently, with the success of Transformer model in speech community [1, 2], the Transformer [3, 4] and its variants [5] have successfully achieved superior performance on this task.”