mlsanityData to 7 Oct 2026

A New Training Pipeline for an Improved Neural Transducer

arXiv 2005.09319 ↗First posted 19 May 2020eess.AS cs.LG cs.NEcited by 34

Not on the map. Why?

Abstract (arXiv)

The RNN transducer is a promising end-to-end model candidate. We compare the original training criterion with the full marginalization over all alignments, to the commonly used maximum approximation, which simplifies, improves and speeds up our training. We also generalize from the original neural network model and study more powerful models, made possible due to the maximum approximation. We further generalize the output label topology to cover RNN-T, RNA and CTC. We perform several studies among all these aspects, including a study on the effect of external alignments. We find that the transducer model generalizes much better on longer sequences than the attention model. Our final transducer model outperforms our attention model on Switchboard 300h by over 6% relative WER.

Key facts. Cited by 34 papers in the map (data through 7 Oct 2026). First posted on arXiv 19 May 2020.

Builds on

No reference list yet.

Built on by

Cited by year

Citing papers per year: 2021: 10; 2022: 12; 2023: 9; 2024: 1; 2025: 2.

34 papers.

Sort:

  1. 2023End-to-End Speech Recognition: A Surveycited by 127
  2. 2021Advancing RNN Transducer Technology for Speech Recognitioncited by 44
  3. 2021Librispeech Transducer Model with Internal Language Model Prior Correctioncited by 26
  4. 2022JOIST: A Joint Speech and Text Streaming Model For ASRcited by 24
  5. 20224-bit Conformer with Native Quantization Aware Training for Speech Recognitioncited by 22
  6. 2021On Addressing Practical Challenges for RNN-Transducercited by 19
  7. 2022Towards Contextual Spelling Correction for Customization of End-to-end Speech Recognition Systemscited by 17
  8. 2022Modular Hybrid Autoregressive Transducercited by 17
  9. 2022Turn-Taking Prediction for Natural Conversational Speechcited by 15
  10. 2022Efficient Training of Neural Transducer for Speech Recognitioncited by 14
  11. 2022Bayes risk CTC: Controllable CTC alignment in Sequence-to-Sequence taskscited by 12
  12. 2021Acoustic Data-Driven Subword Modeling for End-to-End Speech Recognitioncited by 11
22 more
  1. 2021A Light-weight contextual spelling correction model for customizing transducer-based speech recognition systemscited by 11
  2. 2022Improving the fusion of acoustic and text representations in RNN-Tcited by 11
  3. 2022A Unified Cascaded Encoder ASR Model for Dynamic Model Sizescited by 11
  4. 2021Input Length Matters: Improving RNN-T and MWER Training for Long-form Telephony Speech Recognitioncited by 10
  5. 2023Massively Multilingual Shallow Fusion with Large Language Modelscited by 10
  6. 2022A Language Agnostic Multilingual Streaming On-Device ASR Systemcited by 9
  7. 2023JEIT: Joint End-to-End Model and Internal Language Model Training for Speech Recognitioncited by 9
  8. 2021On lattice-free boosted MMI training of HMM and CTC-based full-context ASR modelscited by 7
  9. 2023Chunked Attention-based Encoder-Decoder Model for Streaming Speech Recognitioncited by 7
  10. 2021A study of latent monotonic attention variantscited by 5
  11. 2021Equivalence of Segmental and Neural Transducer Modeling: A Proof of Conceptcited by 5
  12. 2023A CTC Alignment-based Non-autoregressive Transformer for End-to-end Automatic Speech Recognitioncited by 5
  13. 2021Sequence Transduction with Graph-based Supervisioncited by 4
  14. 2022Monotonic segmental attention for automatic speech recognitioncited by 4
  15. 2022Streaming Intended Query Detection using E2E Modeling for Continued Conversationcited by 3
  16. 2023Lego-Features: Exporting modular encoder features for streaming and deliberation ASRcited by 3
  17. 2024Text Injection for Neural Contextual Biasingcited by 3
  18. 2023Bayes Risk Transducer: Transducer with Controllable Alignment Predictioncited by 1
  19. 2023Practical Conformer: Optimizing size, speed and flops of Conformer for on-Device and cloud ASRcited by 0
  20. 2023RAND: Robustness Aware Norm Decay For Quantized Seq2seq Modelscited by 0
  21. 2025A Differentiable Alignment Framework for Sequence-to-Sequence Modeling via Optimal Transportcited by 0
  22. 2025Audio-3DVG: Unified Audio -- Point Cloud Fusion for 3D Visual Groundingcited by 0