mlsanityData to 7 Oct 2026

Back from the future: bidirectional CTC decoding using future information in speech recognition

arXiv 2110.03326 ↗First posted 7 Oct 2021cs.CL cs.SD eess.AScited by 2

Not on the map. Why?

Abstract (arXiv)

In this paper, we propose a simple but effective method to decode the output of Connectionist Temporal Classifier (CTC) model using a bi-directional neural language model. The bidirectional language model uses the future as well as the past information in order to predict the next output in the sequence. The proposed method based on bi-directional beam search takes advantage of the CTC greedy decoding output to represent the noisy future information. Experiments on the Librispeechdataset demonstrate the superiority of our proposed method compared to baselines using unidirectional decoding. In particular, the boost inaccuracy is most apparent at the start of a sequence which is the most erroneous part for existing systems based on unidirectional decoding.

Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Builds on 15 of its 26 references that are in the map. First posted on arXiv 7 Oct 2021.

Builds on

15 of 26 references.

References · 15

  1. 2020Conformer: Convolution-augmented Transformer for Speech Recognitioncited by 1,838
  2. 2019SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognitioncited by 1,355
  3. 2015Deep Speech 2: End-to-End Speech Recognition in English and Mandarincited by 449
  4. 2017Regularizing and Optimizing LSTM Language Modelscited by 205
  5. 2017State-of-the-art Speech Recognition With Sequence-to-Sequence Modelscited by 196
  6. 2015End-to-End Attention-based Large Vocabulary Speech Recognitioncited by 148
  7. 2014Long Short-Term Memory Based Recurrent Neural Network Architectures for Large Vocabulary Speech Recognitioncited by 111
  8. 2018Exploring Architectures, Data and Units For Streaming End-to-End Speech Recognition with RNN-Transducercited by 95
  9. 2014First-Pass Large Vocabulary Continuous Speech Recognition using Bi-Directional Recurrent DNNscited by 27
  10. 2020Improving RNN Transducer Based ASR with Auxiliary Taskscited by 27
  11. 2021Librispeech Transducer Model with Internal Language Model Prior Correctioncited by 26
  12. 2018Asynchronous Bidirectional Decoding for Neural Machine Translationcited by 21
3 more
  1. 2018Fully Convolutional Speech Recognitioncited by 19
  2. 2018Improved training of neural trans-dimensional random field language models with dynamic noise-contrastive estimationcited by 5
  3. 2017Future Word Contexts in Neural Network Language Modelscited by 2

Built on by

Cited by year

Citing papers per year: 2022: 1; 2023: 1.

2 papers.

  1. 2023Iterative Shallow Fusion of Backward Language Model for End-to-End Speech Recognitioncited by 1
  2. 2022Korean Tokenization for Beam Search Rescoring in Speech Recognitioncited by 0