mlsanityData to 7 Oct 2026

A Density Ratio Approach to Language Model Fusion in End-To-End Automatic Speech Recognition

arXiv 2002.11268 ↗First posted 26 Feb 2020eess.AS cs.CL cs.SDcited by 73

Not on the map. Why?

Abstract (arXiv)

This article describes a density ratio approach to integrating external Language Models (LMs) into end-to-end models for Automatic Speech Recognition (ASR). Applied to a Recurrent Neural Network Transducer (RNN-T) ASR model trained on a given domain, a matched in-domain RNN-LM, and a target domain RNN-LM, the proposed method uses Bayes' Rule to define RNN-T posteriors for the target domain, in a manner directly analogous to the classic hybrid model for ASR based on Deep Neural Networks (DNNs) or LSTMs in the Hidden Markov Model (HMM) framework (Bourlard & Morgan, 1994). The proposed approach is evaluated in cross-domain and limited-data scenarios, for which a significant amount of target domain text data is used for LM training, but only limited (or no) {audio, transcript} training data pairs are used to train the RNN-T. Specifically, an RNN-T model trained on paired audio & transcript data from YouTube is evaluated for its ability to generalize to Voice Search data. The Density Ratio method was found to consistently outperform the dominant approach to LM and end-to-end ASR integration, Shallow Fusion.

Key facts. Cited by 73 papers in the map (data through 7 Oct 2026). First posted on arXiv 26 Feb 2020.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 16; 2022: 22; 2023: 24; 2024: 6; 2025: 3; 2026: 2.

73 papers.

Sort:

By area:

  1. 2021Recent Advances in End-to-End Automatic Speech Recognitioncited by 138
  2. 2023End-to-End Speech Recognition: A Surveycited by 127
  3. 2021Contextualized Streaming End-to-End Speech Recognition with Trie-Based Deep Biasing and Shallow Fusioncited by 100
  4. 2023Prompting Large Language Models for Zero-Shot Domain Adaptation in Speech Recognitioncited by 48
  5. 2021Advancing RNN Transducer Technology for Speech Recognitioncited by 44
  6. 2021Internal Language Model Training for Domain-Adaptive End-to-End Speech Recognitioncited by 37
  7. 2021SynthASR: Unlocking Synthetic Data for Speech Recognitioncited by 34
  8. 2021Factorized Neural Transducer for Efficient Language Model Adaptationcited by 31
  9. 2021On the limit of English conversational speech recognitioncited by 30
  10. 2021Librispeech Transducer Model with Internal Language Model Prior Correctioncited by 26
  11. 2021Investigating Methods to Improve Language Model Integration for Attention-based Encoder-Decoder ASR Modelscited by 25
  12. 2024ASR Error Correction using Large Language Modelscited by 23
38 more
  1. 2023An Overview on Language Models: Recent Developments and Outlookcited by 22
  2. 2022Integrating Text Inputs For Training and Adapting RNN Transducer ASR Modelscited by 20
  3. 2021Minimum Word Error Rate Training with Language Model Fusion for End-to-End Speech Recognitioncited by 19
  4. 2021Internal Language Model Adaptation with Text-Only Data for End-to-End Speech Recognitioncited by 17
  5. 2021On Language Model Integration for RNN Transducer based Speech Recognitioncited by 17
  6. 2022Tree-constrained Pointer Generator with Graph Neural Network Encodings for Contextual Speech Recognitioncited by 17
  7. 2022Modular Hybrid Autoregressive Transducercited by 17
  8. 2022Efficient Training of Neural Transducer for Speech Recognitioncited by 14
  9. 2023Large-scale Language Model Rescoring on Long-form Datacited by 14
  10. 2022Improving Rare Word Recognition with LM-aware MWER Trainingcited by 12
  11. 2022Improving the fusion of acoustic and text representations in RNN-Tcited by 11
  12. 2022Minimising Biasing Word Errors for Contextual ASR with the Tree-Constrained Pointer Generatorcited by 11
  13. 2022Fast and accurate factorized neural transducer for text adaption of end-to-end speech recognition modelscited by 11
  14. 2022Internal Language Model Estimation Through Explicit Context Vector Learning for Attention-based Encoder-decoder ASRcited by 9
  15. 2023JEIT: Joint End-to-End Model and Internal Language Model Training for Speech Recognitioncited by 9
  16. 2025When Large Language Models Meet Speech: A Survey on Integration Approachescited by 9
  17. 2021Comparing the Benefit of Synthetic Training Data for Various Automatic Speech Recognition Architecturescited by 8
  18. 2023Approximate Nearest Neighbour Phrase Mining for Contextual Speech Recognitioncited by 8
  19. 2024SALSA: Speedy ASR-LLM Synchronous Aggregationcited by 8
  20. 2021Have best of both worlds: two-pass hybrid and E2E cascading framework for speech recognitioncited by 7
  21. 2022USTED: Improving ASR with a Unified Speech and Text Encoder-Decodercited by 7
  22. 2025NGPU-LM: GPU-Accelerated N-Gram Language Model for Context-Biasing in Greedy ASR Decodingcited by 7
  23. 2023Acoustic Model Fusion for End-to-end Speech Recognitioncited by 6
  24. 2023Personalization of CTC-based End-to-End Speech Recognition Using Pronunciation-Driven Subword Tokenizationcited by 6
  25. 2021Adapting Document-Grounded Dialog Systems to Spoken Conversations using Data Augmentation and a Noisy Channel Modelcited by 5
  26. 2022A Likelihood Ratio based Domain Adaptation Method for E2E Modelscited by 5
  27. 2022Adaptive Discounting of Implicit Language Models in RNN-Transducerscited by 5
  28. 2022Internal Language Model Estimation based Language Model Fusion for Cross-Domain Code-Switching Speech Recognitioncited by 5
  29. 2022Relaxed Attention for Transformer Modelscited by 5
  30. 2022Optimizing Bilingual Neural Transducer with Synthetic Code-switching Text Generationcited by 5
  31. 2022Joint Audio/Text Training for Transformer Rescorer of Streaming Speech Recognitioncited by 5
  32. 2023Text-only Domain Adaptation using Unified Speech-Text Representation in Transducercited by 5
  33. 2024Revisiting ASR Error Correction with Specialized Modelscited by 5
  34. 2024Label-Synchronous Neural Transducer for E2E Simultaneous Speech Translationcited by 5
  35. 2025GEC-RAG: Improving Generative Error Correction via Retrieval-Augmented Generation for Automatic Speech Recognition Systemscited by 5
  36. 2023On-the-fly Text Retrieval for End-to-End ASR Adaptationcited by 4
  37. 2023Improving RNN-Transducers with Acoustic LookAheadcited by 4
  38. 2023Adapting an Unadaptable ASR Systemcited by 3

50 most cited of 73 All 73 (JSON)