Explicit word error minimization in n-best list rescoring

We show that the standard hypothesis scoring paradigm used in maximum-likelihood-based speech recognition systems is not optimal with regard to minimizing the word error rate, the commonly used performance metric in speech recognition. This can lead to sub-optimal performance, especially in high-error-rate environments where word error and sentence error are not necessarily monotonically related. To address this discrepancy, we developed a new algorithm that explicitly minimizes expected word error for recognition hypotheses. First, we approximate the posterior hypothesis probabilities using N-best lists. We then compute the expected word error for each hypothesis with respect to the posterior distribution, and choose the hypothesis with the lowest error. Experiments show improved recognition rates on two spontaneous speech corpora.

• Publication year

  1997

• Venue

  EUROSPEECH

• Publication date

  1997-09-22

• Fields of study

  Computer Science

• Identifiers
  DOI 10.21437/Eurospeech.1997-68
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Neural-network based measures of confidence for word recognition

  1997cited by this paper
• LVCSR log-likelihood ratio scoring for keyword spotting

  1995cited by this paper
• Combining Linguistic and Statistical Knowledge Sources in Natural-Language Processing for ATIS

  1995cited by this paper
• SWITCHBOARD: telephone speech corpus for research and development

  1992cited by this paper
• Integration of Diverse Recognition Methodologies Through Reevaluation of N-Best Sentence Hypotheses

  1991cited by this paper
• A Maximum Likelihood Approach to Continuous Speech Recognition

  1983cited by this paper
• Pattern classification and scene analysis

  1974cited by this paper
• Pattern classification and scene analysis

  1974cited by this paper

• Minimum Bayes Risk Decoding for Error Span Detection in Reference-Free Automatic Machine Translation Evaluation

  2025cites this paper
• Large Language Models for Dysfluency Detection in Stuttered Speech

  2024cites this paper
• Unveiling the Power of Source: Source-based Minimum Bayes Risk Decoding for Neural Machine Translation

  2024cites this paper
• Arabic Speech Recognition: Advancement and Challenges

  2024cites this paper
• Dialect Adaptation and Data Augmentation for Low-Resource ASR: Taltech Systems for the Madasr 2023 Challenge

  2023cites this paper
• Epsilon Sampling Rocks: Investigating Sampling Strategies for Minimum Bayes Risk Decoding for Machine Translation

  2023cites this paper
• Modelling Inter-Rater Uncertainty in Spoken Language Assessment

  2023cites this paper
• It’s MBR All the Way Down: Modern Generation Techniques Through the Lens of Minimum Bayes Risk

  2023cites this paper
• On The Diversity of ASR Hypotheses In Spoken Language Understanding

  2022cites this paper
• LET-Decoder: A WFST-Based Lazy-Evaluation Token-Group Decoder With Exact Lattice Generation

  2021cites this paper
• High Quality Rather than High Model Probability: Minimum Bayes Risk Decoding with Neural Metrics

  2021cites this paper
• Minimum Bayes Risk Decoding with Neural Metrics of Translation Quality

  2021cites this paper
• Combining Hybrid and End-to-End Approaches for the OpenASR20 Challenge

  2021cites this paper
• Sampling-Based Minimum Bayes Risk Decoding for Neural Machine Translation

  2021cites this paper
• Class-Selective Mini-Batching and Multitask Learning for Visual Relationship Recognition

  2021cites this paper
• An Asynchronous WFST-Based Decoder for Automatic Speech Recognition

  2021cites this paper
• Sampling-Based Approximations to Minimum Bayes Risk Decoding for Neural Machine Translation

  2021cites this paper
• ASAPP-ASR: Multistream CNN and Self-Attentive SRU for SOTA Speech Recognition

  2020cites this paper
• Large Margin Training for Attention Based End-to-End Speech Recognition

  2019cites this paper
• Word Importance Modeling to Enhance Captions Generated by Automatic Speech Recognition for Deaf and Hard of Hearing Users

  2019influential citation
• Contextual language understanding Thoughts on Machine Learning in Natural Language Processing

  2019cites this paper
• Image-speech combination for interactive computer assisted transcription of handwritten documents

  2019influential citation
• Adversarial training and decoding strategies for end-to-end neural conversation models

  2019cites this paper
• Ensemble generation and compression for speech recognition

  2019cites this paper
• Watch, Listen and Tell: Multi-Modal Weakly Supervised Dense Event Captioning

  2019cites this paper
• Rescoring of N-Best Hypotheses Using Top-Down Selective Attention for Automatic Speech Recognition

  2018cites this paper
• Automatic Speech Assessment for People with Aphasia Using TDNN-BLSTM with Multi-Task Learning

  2018cites this paper
• Advances on the Transcription of Historical Manuscripts based on Multimodality, Interactivity and Crowdsourcing

  2018cites this paper
• Domain adaptation for statistical machine translation and neural machine translation

  2017cites this paper
• Use of Knowledge Graph in Rescoring the N-Best List in Automatic Speech Recognition

  2017cites this paper
• Early and late integration of audio features for automatic video description

  2017cites this paper
• Étude sur les représentations continues de mots appliquées à la détection automatique des erreurs de reconnaissance de la parole. (A study of continuous word representations applied to the automatic detection of speech recognition errors)

  2017cites this paper
• Sequence Adversarial Training and Minimum Bayes Risk Decoding for End-to-end Neural Conversation Models

  2017cites this paper
• Étude sur les représentations continues de mots appliquées à la détection automatique des erreurs de reconnaissance de la parole

  2017cites this paper
• Evaluation of Language Models over Croatian Newspaper Texts

  2017cites this paper
• Effect of Speech Recognition Errors on Text Understandability for People who are Deaf or Hard of Hearing

  2016influential citation
• Computing the Expected Edit Distance from a String to a PFA

  2016cites this paper
• Adaptive Boosting for Automatic Speech Recognition

  2016cites this paper
• Improvements in language and translation modeling

  2016cites this paper
• Combination of multiple acoustic models with unsupervised adaptation for lecture speech transcription

  2016cites this paper
• Information Fusion Approaches for Distant Speech Recognition in a Multi-microphone Setting

  2016cites this paper
• Word segmentation and pronunciation extraction from phoneme sequences through cross-lingual word-to-phoneme alignment

  2016cites this paper
• Graphical Models for Peptide Identification of Tandem Mass Spectra

  2016cites this paper
• A survey on the application of recurrent neural networks to statistical language modeling

  2015cites this paper
• Hypotheses ranking and state tracking for a multi-domain dialog system using multiple ASR alternates

  2015cites this paper
• Modeling of Slovak Language for Broadcast News Transcription

  2015cites this paper
• Improving Speech Recognizer Performance in a Dialog System Using N-best Hypotheses Reranking

  2015cites this paper
• Automatic Speech Recognition for Low-resource Languages and Accents Using Multilingual and Crosslingual Information

  2014cites this paper
• Towards automatic speech recognition without pronunciation dictionary, transcribed speech and text resources in the target language using cross-lingual word-to-phoneme alignment

  2014cites this paper
• Ensemble Methods for Historical Machine-Printed Document Recognition

  2014cites this paper
• Language Models With Meta-information

  2014cites this paper
• Robust ASR in Reverberant Environments using Temporal Cepstrum Smoothing for Speech Enhancement and an Amplitude Modulation Filterbank for Feature Extraction

  2014cites this paper
• Recent advances in the statistical modeling of the Slovak language

  2014cites this paper
• Channel selection and reverberation-robust automatic speech recognition

  2013cites this paper
• Pronunciation Extraction from Phoneme Sequences through Cross-Lingual Word-to-Phoneme Alignment

  2013cites this paper
• Channel selection using n-best hypothesis for multi-microphone ASR

  2013cites this paper
• Revisiting hybrid and GMM-HMM system combination techniques

  2013cites this paper
• Exploiting the succeeding words in recurrent neural network language models

  2013cites this paper
• Semantic parsing using word confusion networks with conditional random fields

  2013cites this paper
• WFST Enabled Solutions to ASR Problems: Beyond HMM Decoding

  2012cites this paper
• Integration of multiple acoustic and language models for improved Hindi speech recognition system

  2012cites this paper
• Improving the Readability of ASR Results for Lectures Using Multiple Hypotheses and Sentence-Level Knowledge

  2012cites this paper
• 音声ドキュメントの音声認識,整形,要約に関する研究

  2012cites this paper
• Aportaciones al modelado conexionista de lenguaje y su aplicación al reconocimiento de secuencias y traducción automática

  2012cites this paper
• Bidirectional Language Model for Handwriting Recognition

  2012cites this paper
• Topic Mining based on Word Posterior Probability in Spoken Document

  2011cites this paper
• Aniterative approach to Bayes risk decoding and system combination

  2011cites this paper
• Optimal Search for Minimum Error Rate Training

  2011cites this paper
• Bayes risk decoding and its application to system combination

  2011cites this paper
• Investigation of improved approaches to bayes risk decoding

  2011cites this paper
• Minimum Bayes risk discriminative language models for Arabic speech recognition

  2011cites this paper
• Minimum Bayes Risk decoding and system combination based on a recursion for edit distance

  2011cites this paper
• Using N-Best Lists and Confusion Networks for Meeting Summarization

  2011cites this paper
• Search and decoding strategies for complex lexical modeling in lvcsr

  2011cites this paper
• Fusion de données audio-visuelles pour l'interaction Homme-Robot

  2010cites this paper
• Different Evaluation Approaches of Confusion Network in Chinese Spoken Classification

  2010cites this paper
• An improved minimum word error approachto lattice rescoring and system combination

  2010cites this paper
• An improved consensus-like method for Minimum Bayes Risk decoding and lattice combination

  2010cites this paper
• Morpho-syntactic post-processing of N-best lists for improved French automatic speech recognition

  2010cites this paper
• Improving the Readability of Class Lecture Automatic Speech Recognition Results using Multiple Hypotheses

  2010cites this paper
• Using Confusion Networks for Speech Summarization

  2010cites this paper
• Robust machine translation for multi-domain tasks

  2010cites this paper
• Multimodal Processing and Interaction, Audio, Video, Text

  2010cites this paper
• Automatic extractive summarization on meeting corpus

  2010cites this paper
• Pseudo Conditional Random Fields: Joint Training Approach to Segmenting and Labeling Sequence Data

  2010cites this paper
• Minimum hypothesis phone error as a decoding method for speech recognition

  2009influential citation
• A study on integrating acoustic-phonetic information into lattice rescoring for automatic speech recognition

  2009cites this paper
• A Feature Selection Approach for Automatic Music Genre Classification

  2009cites this paper
• Iterative decoding: A novel re-scoring framework for confusion networks

  2009cites this paper
• Morphosyntactic Resources for Automatic Speech Recognition

  2008cites this paper
• Minimum Bayes-Risk decoding with presumedword significance for speech based information retrieval

  2008cites this paper
• Development of SRI’s 1997 Broadcast News Transcription System

  2008cites this paper
• Word/sub-word lattices decomposition and combination for speech recognition

  2008cites this paper
• Efficient Error Correction for Speech Recognition Systems using Constrained Re-recognition

  2008cites this paper
• Construction et stratégie d'exploitation des réseaux de confusion en lien avec le contexte applicatif de la compréhension de la parole. (Confusion networks : construction algorithms and Spoken Language Understanding decision strategies in real applications)

  2008cites this paper
• Efficient error correction for speech systems using constrained re-recognition

  2008cites this paper
• Exploiting non-linear probabilistic models in natural language parsing and reranking

  2008cites this paper
• Automatic sentence structure annotation for spoken language processing

  2008cites this paper
• Toward the Integration of Natural Language Processing and Automatic Speech Recognition: Using Morpho-Syntax and Pragmatics for Transcription

  2008cites this paper
• Confidence based multiple classifier fusion in speaker verification

  2008cites this paper