mlsanityData to 7 Oct 2026

Approximate Nearest Neighbour Phrase Mining for Contextual Speech Recognition

arXiv 2304.08862 ↗First posted 18 Apr 2023cs.CL eess.AScited by 8

Not on the map. Why?

Abstract (arXiv)

This paper presents an extension to train end-to-end Context-Aware Transformer Transducer ( CATT ) models by using a simple, yet efficient method of mining hard negative phrases from the latent space of the context encoder. During training, given a reference query, we mine a number of similar phrases using approximate nearest neighbour search. These sampled phrases are then used as negative examples in the context list alongside random and ground truth contextual information. By including approximate nearest neighbour phrases (ANN-P) in the context list, we encourage the learned representation to disambiguate between similar, but not identical, biasing phrases. This improves biasing accuracy when there are several similar phrases in the biasing inventory. We carry out experiments in a large-scale data regime obtaining up to 7% relative word error rate reductions for the contextual portion of test data. We also extend and evaluate CATT approach in streaming applications.

Key facts. Cited by 8 papers in the map (data through 7 Oct 2026). Builds on 22 of its 33 references that are in the map. First posted on arXiv 18 Apr 2023.

Builds on

22 of 33 references.

References · 22

  1. 2017Attention Is All You Needcited by 59,932
  2. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  3. 2020Conformer: Convolution-augmented Transformer for Speech Recognitioncited by 1,838
  4. 2020Transformer Transducer: A Streamable Speech Recognition Model with Transformer Encoders and RNN-T Losscited by 232
  5. 2018Deep context: end-to-end contextual speech recognitioncited by 136
  6. 2020Emformer: Efficient Memory Transformer Based Acoustic Model For Low Latency Streaming Speech Recognitioncited by 111
  7. 2020Developing Real-time Streaming Transformer Transducer for Speech Recognition on Large-scale Datasetcited by 94
  8. 2020Contextual RNN-T For Open Domain ASRcited by 85
  9. 2020Internal Language Model Estimation for Domain-Adaptive End-to-End Speech Recognitioncited by 80
  10. 2020A Density Ratio Approach to Language Model Fusion in End-To-End Automatic Speech Recognitioncited by 73
  11. 2020FastEmit: Low-latency Streaming ASR with Sequence-level Emission Regularizationcited by 66
  12. 2020Deep Shallow Fusion for RNN-T Personalizationcited by 64
10 more
  1. 2019A spelling correction model for end-to-end speech recognitioncited by 63
  2. 2022Contextual Adapters for Personalized Speech Recognition in Neural Transducerscited by 57
  3. 2021Context-Aware Transformer Transducer for Speech Recognitioncited by 56
  4. 2021Tree-constrained Pointer Generator for End-to-end Contextual Speech Recognitioncited by 32
  5. 2018Contextual Speech Recognition with Difficult Negative Training Examplescited by 28
  6. 2021Fast Contextual Adaptation with Neural Associative Memory for On-Device Personalized Speech Recognitioncited by 25
  7. 2020Transformer Transducer: One Model Unifying Streaming and Non-streaming Speech Recognitioncited by 23
  8. 2022Variable Attention Masking for Configurable Transformer Transducer Speech Recognitioncited by 14
  9. 2017No Need for a Lexicon? Evaluating the Value of the Pronunciation Lexica in End-to-End Modelscited by 13
  10. 2022Optimizing Bilingual Neural Transducer with Synthetic Code-switching Text Generationcited by 5

Built on by

Cited by year

Citing papers per year: 2023: 3; 2024: 4; 2025: 1.

8 papers.

  1. 2024Contextualization of ASR with LLM using phonetic retrieval-based augmentationcited by 10
  2. 2023Retrieve and Copy: Scaling ASR Personalization to Large Catalogscited by 7
  3. 2023Multimodal Data and Resource Efficient Device-Directed Speech Detection with Large Foundation Modelscited by 5
  4. 2024A Multimodal Approach to Device-Directed Speech Detection with Large Language Modelscited by 5
  5. 2024Locality enhanced dynamic biasing and sampling strategies for contextual ASRcited by 3
  6. 2024Optimizing Contextual Speech Recognition Using Vector Quantization for Efficient Retrievalcited by 3
  7. 2025Enhancing the Robustness of Contextual ASR to Varying Biasing Information Volumes Through Purified Semantic Correlation Joint Modelingcited by 1
  8. 2023Wiki-En-ASR-Adapt: Large-scale synthetic dataset for English ASR Customizationcited by 0