Error detection and accuracy estimation in automatic speech recognition using deep bidirectional recurrent neural networks

A Ogawa, T Hori - Speech Communication, 2017 - Elsevier
Recurrent neural networks (RNNs) have recently been applied as the classifiers for
sequential labeling problems. In this paper, deep bidirectional RNNs (DBRNNs) are applied …

Estimating confidence scores on ASR results using recurrent neural networks

K Kalgaonkar, C Liu, Y Gong… - 2015 IEEE International …, 2015 - ieeexplore.ieee.org
In this paper we present a confidence estimation system using recurrent neural networks
(RNN) and compare it to a traditional multilayered perception (MLP) based system. The …

Analyzing uncertainties in speech recognition using dropout

A Vyas, P Dighe, S Tong… - ICASSP 2019-2019 IEEE …, 2019 - ieeexplore.ieee.org
The performance of Automatic Speech Recognition (ASR) systems is often measured using
Word Error Rates (WER) which requires time-consuming and expensive manually …

[PDF][PDF] Combining information sources for confidence estimation with CRF models

MS Seigel, PC Woodland - Twelfth Annual Conference of the …, 2011 - isca-archive.org
Obtaining accurate confidence measures for automatic speech recognition (ASR)
transcriptions is an important task which stands to benefit from the use of multiple …

ASR error detection and recognition rate estimation using deep bidirectional recurrent neural networks

A Ogawa, T Hori - … on Acoustics, Speech and Signal Processing …, 2015 - ieeexplore.ieee.org
Recurrent neural networks (RNNs) have recently been applied as the classifiers for
sequential labeling problems. In this paper, deep bidirectional RNNs (DBRNNs) are applied …

Enhancing lexical cohesion measure with confidence measures, semantic relations and language model interpolation for multimedia spoken content topic …

C Guinaudeau, G Gravier, P Sébillot - Computer Speech & Language, 2012 - Elsevier
Transcript-based topic segmentation of TV programs faces several difficulties arising from
transcription errors, from the presence of potentially short segments and from the limited …

System-independent asr error detection and classification using recurrent neural network

R Errattahi, AEL Hannani, T Hain… - Computer Speech & …, 2019 - Elsevier
This paper addresses errors in continuous Automatic Speech Recognition (ASR) in two
stages: error detection and error type classification. Unlike the majority of research in this …

Estimating speech recognition accuracy based on error type classification

A Ogawa, T Hori, A Nakamura - IEEE/ACM Transactions on …, 2016 - ieeexplore.ieee.org
Methods for estimating the speech recognition accuracy without using manually transcribed
references are beneficial to the research and development of automatic speech recognition …

Speaker-adapted confidence measures for speech recognition of video lectures

I Sanchez-Cortina, J Andrés-Ferrer, A Sanchis… - Computer Speech & …, 2016 - Elsevier
Automatic speech recognition applications can benefit from a confidence measure (CM) to
predict the reliability of the output. Previous works showed that a word-dependent naïve …

An empirical study of software exceptions in the field using search logs

F Hassan, C Bansal, N Nagappan… - Proceedings of the 14th …, 2020 - dl.acm.org
Background: Software engineers spend a substantial amount of time using Web search to
accomplish software engineering tasks. Such search tasks include finding code snippets …