mlsanityData to 14 Oct 2026

Deformable TDNN with adaptive receptive fields for speech recognition

arXiv 2104.14791 ↗First posted 30 Apr 2021eess.AS cs.CL cs.SDcited by 2

Not on the map. Why?

Abstract (arXiv)

Time Delay Neural Networks (TDNNs) are widely used in both DNN-HMM based hybrid speech recognition systems and recent end-to-end systems. Nevertheless, the receptive fields of TDNNs are limited and fixed, which is not desirable for tasks like speech recognition, where the temporal dynamics of speech are varied and affected by many factors. This paper proposes to use deformable TDNNs for adaptive temporal dynamics modeling in end-to-end speech recognition. Inspired by deformable ConvNets, deformable TDNNs augment the temporal sampling locations with additional offsets and learn the offsets automatically based on the ASR criterion, without additional supervision. Experiments show that deformable TDNNs obtain state-of-the-art results on WSJ benchmarks (1.42%/3.45% WER on WSJ eval92/dev93 respectively), outperforming standard TDNNs significantly. Furthermore, we propose the latency control mechanism for deformable TDNNs, which enables deformable TDNNs to do streaming ASR without accuracy degradation.

Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Builds on 16 of its 32 references that are in the map. First posted on arXiv 30 Apr 2021.

Builds on

16 of 32 references.

References · 16

  1. 2017Attention Is All You Needcited by 60,361
  2. 2014Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translationcited by 4,057
  3. 2017Deformable Convolutional Networkscited by 1,488
  4. 2019SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognitioncited by 1,380
  5. 2018ESPnet: End-to-End Speech Processing Toolkitcited by 558
  6. 2019On the Relationship between Self-Attention and Convolutional Layerscited by 300
  7. 2019Jasper: An End-to-End Convolutional Neural Acoustic Modelcited by 76
  8. 2019Self-Attention Aligner: A Latency-Control End-to-End Model for ASR Using Self-Attention Network and Chunk-Hoppingcited by 33
  9. 2019Espresso: A Fast End-to-end Neural Speech Recognition Toolkitcited by 31
  10. 2020Transformer-based Online CTC/attention End-to-End Speech Recognition Architecturecited by 28
  11. 2018Fully Convolutional Speech Recognitioncited by 19
  12. 2020Multistream CNN for Robust Acoustic Modelingcited by 13
4 more
  1. 2020CAT: A CTC-CRF based ASR Toolkit Bridging the Hybrid and the End-to-end Approaches towards Data Efficiency and Low Latencycited by 13
  2. 2019State-of-the-Art Speech Recognition Using Multi-Stream Self-Attention With Dilated 1D Convolutionscited by 11
  3. 2020Efficient Neural Architecture Search for End-to-end Speech Recognition via Straight-Through Gradientscited by 8
  4. 2015Feedforward Sequential Memory Neural Networks without Recurrent Feedbackcited by 2

Built on by

Cited by year

Citing papers per year: 2021: 1; 2022: 1.

2 papers.

  1. 2021Advancing CTC-CRF Based End-to-End Speech Recognition with Wordpieces and Conformerscited by 4
  2. 2022DEFORMER: Coupling Deformed Localized Patterns with Global Context for Robust End-to-end Speech Recognitioncited by 0