mlsanityData to 14 Oct 2026

Improving Neural Biasing for Contextual Speech Recognition by Early Context Injection and Text Perturbation

arXiv 2407.10303 ↗First posted 14 Jul 2024eess.AS cs.CLcited by 11

Not on the map. Why?

Abstract (arXiv)

Existing research suggests that automatic speech recognition (ASR) models can benefit from additional contexts (e.g., contact lists, user specified vocabulary). Rare words and named entities can be better recognized with contexts. In this work, we propose two simple yet effective techniques to improve context-aware ASR models. First, we inject contexts into the encoders at an early stage instead of merely at their last layers. Second, to enforce the model to leverage the contexts during training, we perturb the reference transcription with alternative spellings so that the model learns to rely on the contexts to make correct predictions. On LibriSpeech, our techniques together reduce the rare word error rate by 60% and 25% relatively compared to no biasing and shallow fusion, making the new state-of-the-art performance. On SPGISpeech and a real-world dataset ConEC, our techniques also yield good improvements over the baselines.

Key facts. Cited by 11 papers in the map (data through 14 Oct 2026). Builds on 32 of its 42 references that are in the map. First posted on arXiv 14 Jul 2024.

Builds on

32 of 42 references.

References · 32

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
  2. 2020Conformer: Convolution-augmented Transformer for Speech Recognitioncited by 1,863
  3. 2019SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognitioncited by 1,380
  4. 2018Deep context: end-to-end contextual speech recognitioncited by 137
  5. 2021Contextualized Streaming End-to-End Speech Recognition with Trie-Based Deep Biasing and Shallow Fusioncited by 101
  6. 2020Contextual RNN-T For Open Domain ASRcited by 85
  7. 2023Zipformer: A faster and better encoder for automatic speech recognitioncited by 80
  8. 2021SPGISpeech: 5,000 hours of transcribed financial audio for fully formatted end-to-end speech recognitioncited by 78
  9. 2023SALM: Speech-augmented Language Model with In-context Learning for Speech Recognition and Translationcited by 76
  10. 2020Deep Shallow Fusion for RNN-T Personalizationcited by 64
  11. 2021Earnings-21: A Practical Benchmark for ASR in the Wildcited by 62
  12. 2022Contextual Adapters for Personalized Speech Recognition in Neural Transducerscited by 57
20 more
  1. 2021Context-Aware Transformer Transducer for Speech Recognitioncited by 56
  2. 2020Class LM and word mapping for contextual biasing in End-to-End ASRcited by 42
  3. 2021Instant One-Shot Word-Learning for Context-Specific Neural Sequence-to-Sequence Speech Recognitioncited by 30
  4. 2018Contextual Speech Recognition with Difficult Negative Training Examplescited by 28
  5. 2023End-to-End Speech Recognition Contextualization with Large Language Modelscited by 22
  6. 2023PromptASR for contextualized ASR with controllable stylecited by 19
  7. 2022Towards Contextual Spelling Correction for Customization of End-to-end Speech Recognition Systemscited by 17
  8. 2022Improving Contextual Recognition of Rare Words with an Alternate Spelling Prediction Modelcited by 17
  9. 2024Improving ASR Contextual Biasing with Guided Attentioncited by 17
  10. 2023Phoneme-aware Encoding for Prefix-tree-based Contextual ASRcited by 13
  11. 2019G2G: TTS-Driven Pronunciation Learning for Graphemic Hybrid ASRcited by 11
  12. 2022Minimising Biasing Word Errors for Contextual ASR with the Tree-Constrained Pointer Generatorcited by 11
  13. 2023PROCTER: PROnunciation-aware ConTextual adaptER for personalized speech recognition in neural transducerscited by 10
  14. 2023Contextual Biasing with the Knuth-Morris-Pratt Matching Algorithmcited by 9
  15. 2024Towards ASR Robust Spoken Language Understanding Through In-Context Learning With Word Confusion Networkscited by 8
  16. 2022End-to-end contextual asr based on posterior distribution adaptation for hybrid ctc/attention systemcited by 7
  17. 2023Contextual Biasing of Named-Entities with Large Language Modelscited by 7
  18. 2023Personalization of CTC-based End-to-End Speech Recognition Using Pronunciation-Driven Subword Tokenizationcited by 6
  19. 2022Can Visual Context Improve Automatic Speech Recognition for an Embodied Agent?cited by 3
  20. 2021Improving RNN-T ASR Performance with Date-Time and Location Awarenesscited by 2

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 1; 2025: 8; 2026: 2.

11 papers.

  1. 2025TurboBias: Universal ASR Context-Biasing powered by GPU-accelerated Phrase-Boosting Treecited by 4
  2. 2024Optimizing Contextual Speech Recognition Using Vector Quantization for Efficient Retrievalcited by 3
  3. 2025PMF-CEC: Phoneme-augmented Multimodal Fusion for Context-aware ASR Error Correction with Error-specific Selective Decodingcited by 1
  4. 2025PAC: Pronunciation-Aware Contextualized Large Language Model-based Automatic Speech Recognitioncited by 1
  5. 2025Retrieval Augmented Generation based context discovery for ASRcited by 1
  6. 2025A Neural Model for Contextual Biasing Score Learning and Filteringcited by 1
  7. 2025MGSC: A Multi-granularity Consistency Framework for Robust End-to-end Asrcited by 0
  8. 2025PARCO: Phoneme-Augmented Robust Contextual ASR via Contrastive Entity Disambiguationcited by 0
  9. 2025Peeking Into The Future For Contextual Biasingcited by 0
  10. 2026From Oracle to Noisy Context: Mitigating Contextual Exposure Bias in Speech-LLMscited by 0
  11. 2026TurboBias 2.0: Streaming Context-Biasing for Production-Efficient ASR Systemscited by 0