mlsanityData to 14 Oct 2026

SpellMapper: A non-autoregressive neural spellchecker for ASR customization with candidate retrieval based on n-gram mappings

arXiv 2306.02317 ↗First posted 4 Jun 2023cs.CL cs.AI cs.SDcited by 4

Not on the map. Why?

Abstract (arXiv)

Contextual spelling correction models are an alternative to shallow fusion to improve automatic speech recognition (ASR) quality given user vocabulary. To deal with large user vocabularies, most of these models include candidate retrieval mechanisms, usually based on minimum edit distance between fragments of ASR hypothesis and user phrases. However, the edit-distance approach is slow, non-trainable, and may have low recall as it relies only on common letters. We propose: 1) a novel algorithm for candidate retrieval, based on misspelled n-gram mappings, which gives up to 90% recall with just the top 10 candidates on Spoken Wikipedia; 2) a non-autoregressive neural model based on BERT architecture, where the initial transcript and ten candidates are combined into one input. The experiments on Spoken Wikipedia show 21.4% word error rate improvement compared to a baseline ASR system.

Key facts. Cited by 4 papers in the map (data through 14 Oct 2026). Builds on 17 of its 25 references that are in the map. First posted on arXiv 4 Jun 2023.

Builds on

17 of 25 references.

References · 17

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
  2. 2020HiFi-GAN: Generative Adversarial Networks for Efficient and High Fidelity Speech Synthesiscited by 1,154
  3. 2020FastPitch: Parallel Text-to-speech with Pitch Predictioncited by 151
  4. 2018Deep context: end-to-end contextual speech recognitioncited by 137
  5. 2020Contextual RNN-T For Open Domain ASRcited by 85
  6. 2021SPGISpeech: 5,000 hours of transcribed financial audio for fully formatted end-to-end speech recognitioncited by 78
  7. 2020Deep Shallow Fusion for RNN-T Personalizationcited by 64
  8. 2019A spelling correction model for end-to-end speech recognitioncited by 63
  9. 2019Correction of Automatic Speech Recognition with Transformer Sequence-to-sequence Modelcited by 52
  10. 2019Personalization of End-to-end Speech Recognition On Mobile Devices For Named Entitiescited by 41
  11. 2021Personalization Strategies for End-to-End Speech Recognition Systemscited by 26
  12. 2022Towards Contextual Spelling Correction for Customization of End-to-end Speech Recognition Systemscited by 17
5 more
  1. 2022Improving Contextual Recognition of Rare Words with an Alternate Spelling Prediction Modelcited by 17
  2. 2023Two Stage Contextual Word Filtering for Context bias in Unified Streaming and Non-streaming Transducercited by 12
  3. 2021A Light-weight contextual spelling correction model for customizing transducer-based speech recognition systemscited by 11
  4. 2022UserLibri: A Dataset for ASR Personalization Using Only Textcited by 6
  5. 2021A Toolbox for Construction and Analysis of Speech Datasetscited by 5

Built on by

Cited by year

Citing papers per year: 2023: 1; 2024: 3.

4 papers.

  1. 2024Deferred NAM: Low-latency Top-K Context Injection via Deferred Context Encoding for Non-Streaming ASRcited by 5
  2. 2024Spelling Correction through Rewriting of Non-Autoregressive ASR Latticescited by 1
  3. 2023Wiki-En-ASR-Adapt: Large-scale synthetic dataset for English ASR Customizationcited by 0
  4. 2024High-precision Voice Search Query Correction via Retrievable Speech-text Embedingscited by 0