mlsanityData to 7 Oct 2026

Investigating the Effect of Language Models in Sequence Discriminative Training for Neural Transducers

arXiv 2310.07345 ↗First posted 11 Oct 2023cs.CL cs.SD eess.AScited by 0

Not on the map. Why?

Abstract (arXiv)

In this work, we investigate the effect of language models (LMs) with different context lengths and label units (phoneme vs. word) used in sequence discriminative training for phoneme-based neural transducers. Both lattice-free and N-best-list approaches are examined. For lattice-free methods with phoneme-level LMs, we propose a method to approximate the context history to employ LMs with full-context dependency. This approximation can be extended to arbitrary context length and enables the usage of word-level LMs in lattice-free methods. Moreover, a systematic comparison is conducted across lattice-free and N-best-list-based methods. Experimental results on Librispeech show that using the word-level LM in training outperforms the phoneme-level LM. Besides, we find that the context size of the LM used for probability computation has a limited effect on performance. Moreover, our results reveal the pivotal importance of the hypothesis space quality in sequence discriminative training.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 23 of its 39 references that are in the map. First posted on arXiv 11 Oct 2023.

Builds on

23 of 39 references.

References · 23

  1. 2020Conformer: Convolution-augmented Transformer for Speech Recognitioncited by 1,838
  2. 2015Listen, Attend and Spellcited by 542
  3. 2015End-to-End Attention-based Large Vocabulary Speech Recognitioncited by 148
  4. 2020Hybrid Autoregressive Transducer (hat)cited by 106
  5. 2019Language Modeling with Deep Transformerscited by 96
  6. 2018Exploring Architectures, Data and Units For Streaming End-to-End Speech Recognition with RNN-Transducercited by 95
  7. 2017Minimum Word Error Rate Training for Attention-based Sequence-to-Sequence Modelscited by 68
  8. 2021Internal Language Model Training for Domain-Adaptive End-to-End Speech Recognitioncited by 37
  9. 2020Single headed attention based sequence-to-sequence model for state-of-the-art results on Switchboardcited by 36
  10. 2020Efficient minimum word error rate training of RNN-Transducer for end-to-end speech recognitioncited by 34
  11. 2020Less Is More: Improved RNN-T Decoding Using Limited Label Context and Path Mergingcited by 26
  12. 2017Optimizing expected word error rate via sampling for speech recognitioncited by 21
11 more
  1. 2020Phoneme Based Neural Transducer for Large Vocabulary Speech Recognitioncited by 20
  2. 2021On Language Model Integration for RNN Transducer based Speech Recognitioncited by 17
  3. 2022Efficient Training of Neural Transducer for Speech Recognitioncited by 14
  4. 2020An investigation of phone-based subword units for end-to-end speech recognitioncited by 12
  5. 2021Consistent Training and Decoding For End-to-end Speech Recognition Using Lattice-free MMIcited by 8
  6. 2021On lattice-free boosted MMI training of HMM and CTC-based full-context ASR modelscited by 7
  7. 2019Comparison of Lattice-Free and Lattice-Based Sequence Discriminative Training Criteria for LVCSRcited by 3
  8. 2021Efficient Sequence Training of Attention Models using Approximative Recombinationcited by 3
  9. 2022Integrating Lattice-Free MMI into End-to-End Speech Recognitioncited by 3
  10. 2022Lattice-Free Sequence Discriminative Training for Phoneme-Based Neural Transducerscited by 3
  11. 2023RASR2: The RWTH ASR Toolkit for Generic Sequence-to-sequence Speech Recognitioncited by 3

Built on by

Not cited yet.