A Neural Model for Contextual Biasing Score Learning and Filtering
Not on the map. Why?
Abstract (arXiv)
Contextual biasing improves automatic speech recognition (ASR) by integrating external knowledge, such as user-specific phrases or entities, during decoding. In this work, we use an attention-based biasing decoder to produce scores for candidate phrases based on acoustic information extracted by an ASR encoder, which can be used to filter out unlikely phrases and to calculate bonus for shallow-fusion biasing. We introduce a per-token discriminative objective that encourages higher scores for ground-truth phrases while suppressing distractors. Experiments on the Librispeech biasing benchmark show that our method effectively filters out majority of the candidate phrases, and significantly improves recognition accuracy under different biasing conditions when the scores are used in shallow fusion biasing. Our approach is modular and can be used with any ASR system, and the filtering mechanism can potentially boost performance of other biasing methods.
Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 30 of its 42 references that are in the map. First posted on arXiv 27 Oct 2025.
Builds on
30 of 42 references.
References · 30
- 2017Attention Is All You Needcited by 59,932
- 2018ESPnet: End-to-End Speech Processing Toolkitcited by 554
- 2015Listen, Attend and Spellcited by 542
- 2015Attention-Based Models for Speech Recognitioncited by 400
- 2024Audio Flamingo: A Novel Audio Language Model with Few-Shot Learning and Dialogue Abilitiescited by 183
- 2018Deep context: end-to-end contextual speech recognitioncited by 136
- 2021Contextualized Streaming End-to-End Speech Recognition with Trie-Based Deep Biasing and Shallow Fusioncited by 100
- 2020Contextual RNN-T For Open Domain ASRcited by 85
- 2023SALM: Speech-augmented Language Model with In-context Learning for Speech Recognition and Translationcited by 75
- 2023Generative Speech Recognition Error Correction with Large Language Models and Task-Activating Promptingcited by 71
- 2017Minimum Word Error Rate Training for Attention-based Sequence-to-Sequence Modelscited by 68
- 2023SLM: Bridge the thin gap between speech and text foundation modelscited by 67
18 more
- 2020Deep Shallow Fusion for RNN-T Personalizationcited by 64
- 2022Contextual Adapters for Personalized Speech Recognition in Neural Transducerscited by 57
- 2021Context-Aware Transformer Transducer for Speech Recognitioncited by 56
- 2022Improving End-to-End Contextual Speech Recognition with Fine-Grained Contextual Knowledge Selectioncited by 39
- 2021Tree-constrained Pointer Generator for End-to-end Contextual Speech Recognitioncited by 32
- 2021Fast Contextual Adaptation with Neural Associative Memory for On-Device Personalized Speech Recognitioncited by 25
- 2023End-to-End Speech Recognition Contextualization with Large Language Modelscited by 22
- 2024Contextualized Automatic Speech Recognition with Attention-Based Bias Phrase Boosted Beam Searchcited by 19
- 2024Improving ASR Contextual Biasing with Guided Attentioncited by 17
- 2024Improving Neural Biasing for Contextual Speech Recognition by Early Context Injection and Text Perturbationcited by 11
- 2024Contextualization of ASR with LLM using phonetic retrieval-based augmentationcited by 10
- 2023Contextual Biasing with the Knuth-Morris-Pratt Matching Algorithmcited by 9
- 2024Contextualized End-to-end Automatic Speech Recognition with Intermediate Biasing Losscited by 9
- 2023Contextual Biasing of Named-Entities with Large Language Modelscited by 7
- 2025Delayed Fusion: Integrating Large Language Models into First-Pass Decoding in End-to-end Speech Recognitioncited by 5
- 2024Text Injection for Neural Contextual Biasingcited by 3
- 2024Optimizing Contextual Speech Recognition Using Vector Quantization for Efficient Retrievalcited by 3
- 2024An Effective Context-Balanced Adaptation Approach for Long-Tailed Speech Recognitioncited by 1
Built on by
Citing papers per year: 2026: 1.
1 paper.