mlsanityData to 14 Oct 2026

Global Normalization for Streaming Speech Recognition in a Modular Framework

arXiv 2205.13674 ↗First posted 26 May 2022cs.LG cs.AI cs.CLcited by 10

Not on the map. Why?

Abstract (arXiv)

We introduce the Globally Normalized Autoregressive Transducer (GNAT) for addressing the label bias problem in streaming speech recognition. Our solution admits a tractable exact computation of the denominator for the sequence-level normalization. Through theoretical and empirical results, we demonstrate that by switching to a globally normalized model, the word error rate gap between streaming and non-streaming speech-recognition models can be greatly reduced (by more than 50% on the Librispeech dataset). This model is developed in a modular framework which encompasses all the common neural speech recognition models. The modularity of this framework enables controlled comparison of modelling choices and creation of new models.

Key facts. Cited by 10 papers in the map (data through 14 Oct 2026). Builds on 11 of its 46 references that are in the map. First posted on arXiv 26 May 2022.

Builds on

11 of 46 references.

References · 11

  1. 2017Attention Is All You Needcited by 60,361
  2. 2014Adam: A Method for Stochastic Optimizationcited by 37,240
  3. 2020Conformer: Convolution-augmented Transformer for Speech Recognitioncited by 1,863
  4. 2020Transformer Transducer: A Streamable Speech Recognition Model with Transformer Encoders and RNN-T Losscited by 235
  5. 2020Hybrid Autoregressive Transducer (hat)cited by 108
  6. 2016Wav2Letter: an End-to-End ConvNet-based Speech Recognition Systemcited by 43
  7. 2016Globally Normalized Transition-Based Neural Networkscited by 26
  8. 2019An Empirical Investigation of Global and Local Normalization for Recurrent Neural Sequence Models Using a Continuous Relaxation to Beam Searchcited by 12
  9. 2017End-to-End Neural Segmental Models for Speech Recognitioncited by 4
  10. 2021Advancing CTC-CRF Based End-to-End Speech Recognition with Wordpieces and Conformerscited by 4
  11. 2015Discriminative Segmental Cascades for Feature-Rich Phone Recognitioncited by 1

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2022: 1; 2023: 4; 2024: 2; 2025: 0; 2026: 3.

10 papers.

  1. 2023Powerful and Extensible WFST Framework for RNN-Transducer Lossescited by 7
  2. 2024Preference Optimization for Molecule Synthesis with Conditional Residual Energy-based Modelscited by 6
  3. 2023Exploring Energy-based Language Models with Different Architectures and Training Methods for Speech Recognitioncited by 1
  4. 2022Alignment Entropy Regularizationcited by 0
  5. 2023LAST: Scalable Lattice-Based Speech Modelling in JAXcited by 0
  6. 2023Globally Normalising the Transducer for Streaming Speech Recognitioncited by 0
  7. 2024Towards Maximum Likelihood Training for Transducer-based Streaming Speech Recognitioncited by 0
  8. 2026Do we really need Self-Attention for Streaming Automatic Speech Recognition?cited by 0
  9. 2026SENS-ASR: Semantic Embedding injection in Neural-transducer for Streaming Automatic Speech Recognitioncited by 0
  10. 2026Streaming Speech-to-Text Translation with a SpeechLLMcited by 0