Re-evaluating Minimum Bayes Risk Decoding for Automatic Speech Recognition
Not on the map. Why?
Abstract (arXiv)
Recent work has shown that sample-based Minimum Bayes Risk (MBR) decoding outperforms beam search in text-to-text generation tasks, such as machine translation, text summarization, and image captioning. On the other hand, beam search is the current practice for speech-to-text tasks such as automatic speech recognition (ASR) and Speech Translation (ST). Given that MBR decoding is effective in text-to-text generation tasks, it is reasonable to expect it to also be effective for speech-to-text tasks. In this paper, we evaluate MBR decoding for ASR and ST tasks on English and Japanese using Whisper and its derivative models. We observe that the accuracy of MBR decoding outperforms that of beam search in most of the experimental settings we have evaluated. The results show that MBR decoding is a promising method for offline ASR and ST tasks that require high accuracy. The code is available at https://github.com/CyberAgentAILab/mbr-for-asr
Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Builds on 61 of its 102 references that are in the map. First posted on arXiv 22 Oct 2025.
Builds on
61 of 102 references.
References · 61
- 2017Attention Is All You Needcited by 60,361
- 2019Sentence-BERT: Sentence Embeddings using Siamese BERT-Networkscited by 9,170
- 2022Robust Speech Recognition via Large-Scale Weak Supervisioncited by 4,057
- 2020wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representationscited by 3,905
- 2014Sequence to Sequence Learning with Neural Networkscited by 2,986
- 2019The Curious Case of Neural Text Degenerationcited by 2,916
- 2018A Call for Clarity in Reporting BLEU Scorescited by 1,799
- 2021WebGPT: Browser-assisted question-answering with human feedbackcited by 1,745
- 2019Common Voice: A Massively-Multilingual Speech Corpuscited by 1,384
- 2020BLEURT: Learning Robust Metrics for Text Generationcited by 1,203
- 2020Making Monolingual Sentence Embeddings Multilingual using Knowledge Distillationcited by 593
- 2022FLEURS: Few-shot Learning Evaluation of Universal Representations of Speechcited by 469
38 more
- 2021VoxPopuli: A Large-Scale Multilingual Speech Corpus for Representation Learning, Semi-Supervised Learning and Interpretationcited by 463
- 2015MUSAN: A Music, Speech, and Noise Corpuscited by 425
- 2016Joint CTC-Attention based End-to-End Speech Recognition using Multi-task Learningcited by 257
- 2017Challenges in Data-to-Document Generationcited by 226
- 2020fairseq S2T: Fast Speech-to-Text Modeling with fairseqcited by 176
- 2020Is MAP Decoding All You Need? The Inadequacy of the Mode in Neural Machine Translationcited by 138
- 2023End-to-End Speech Recognition: A Surveycited by 129
- 2022Truncation Sampling as Language Model Desmoothingcited by 123
- 2021High Quality Rather than High Model Probability: Minimum Bayes Risk Decoding with Neural Metricscited by 104
- 2023Distil-Whisper: Robust Knowledge Distillation via Large-Scale Pseudo Labellingcited by 90
- 2019Recognizing long-form speech using streaming end-to-end modelscited by 77
- 2023Generative Speech Recognition Error Correction with Large Language Models and Task-Activating Promptingcited by 72
- 2018What Doubling Tricks Can and Can't Do for Multi-Armed Banditscited by 71
- 2021Sampling-Based Approximations to Minimum Bayes Risk Decoding for Neural Machine Translationcited by 65
- 2019A spelling correction model for end-to-end speech recognitioncited by 63
- 2017JSUT corpus: free large-scale Japanese speech corpus for end-to-end speech synthesiscited by 58
- 2023Epsilon Sampling Rocks: Investigating Sampling Strategies for Minimum Bayes Risk Decoding for Machine Translationcited by 58
- 2020Deliberation Model Based Two-Pass End-to-End Speech Recognitioncited by 57
- 2023Can Generative Large Language Models Perform ASR Error Correction?cited by 57
- 2023Whispering LLaMA: A Cross-Modal Generative Error Correction Framework for Speech Recognitioncited by 56
- 2022Follow the Wisdom of the Crowd: Effective Text Generation via Minimum Bayes Risk Decodingcited by 54
- 2019Correction of Automatic Speech Recognition with Transformer Sequence-to-sequence Modelcited by 52
- 2021FastCorrect: Fast Error Correction with Edit Alignment for Automatic Speech Recognitioncited by 43
- 2019A comparison of end-to-end models for long-form speech recognitioncited by 42
- 2021Semantic Distance: A New Metric for ASR Performance Analysis Towards Spoken Language Understandingcited by 42
- 2023HyPoradise: An Open Baseline for Generative Speech Recognition with Large Language Modelscited by 42
- 2025Evaluation of Best-of-N Sampling Strategies for Language Model Alignmentcited by 39
- 2022RescoreBERT: Discriminative Speech Recognition Rescoring with BERTcited by 34
- 2023Faster Minimum Bayes Risk Decoding with Confidence-based Pruningcited by 34
- 2024Large Language Models are Efficient Learners of Noise-Robust Speech Recognitioncited by 31
- 2021Innovative Bert-based Reranking Language Models for Speech Recognitioncited by 23
- 2024It's Never Too Late: Fusing Acoustic Information into Large Language Models for Automatic Speech Recognitioncited by 22
- 2024Better Instruction-Following Through Minimum Bayes Riskcited by 19
- 2024Findings of the IWSLT 2024 Evaluation Campaigncited by 19
- 2024Mitigating Metric Bias in Minimum Bayes Risk Decodingcited by 16
- 2024Hyperparameter-Free Approach for Faster Minimum Bayes Risk Decodingcited by 15
- 2024Efficient Minimum Bayes Risk Decoding using Low-Rank Matrix Completion Algorithmscited by 13
- 2024Centroid-Based Efficient Minimum Bayes Risk Decodingcited by 12
50 most cited of 61 All 61 (JSON)
Built on by
Citing papers per year: 2025: 1; 2026: 1.
2 papers.