2017/02/06 by Daniele Falavigna, Marco Matassoni, Falavigna, Daniele +7
Computer Science · Mathematics · #Adaptation (eye) #Artificial intelligence #Artificial neural network #Baseline (sea) #Computation and Language (cs.CL) #Computer science #Deep neural networks #Exploit #FOS: Computer and information sciences #Mathematics #Music and Audio Processing #Natural Language Processing Techniques #Oracle #Quality (philosophy) #Sentence #Speech Recognition and Synthesis #Speech recognition #Word (group theory) #Word error rate #cs.CL
paper · pdf · doi:10.48550/arxiv.1702.01714
Computer Speech & Language December 2016
arxiv created 2017/02/06 · openalex publication_date 2017/02/06 · arxiv updated 2017/02/07 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/04
In this paper we propose to exploit the automatic Quality Estimation (QE) of ASR hypotheses to perform the unsupervised adaptation of a deep neural network modeling acoustic probabilities. Our hypothesis is that significant improvements can be achieved by: i)automatically transcribing the evaluation data we are currently trying to recognise, and ii) selecting from it a subset of "good quality" instances based on the word error rate (WER) scores predicted by a QE component. To validate this hypothesis, we run several experiments on the evaluation data sets released for the CHiME-3 challenge. First, we operate in oracle conditions in which manual transcriptions of the evaluation data are available, thus allowing us to compute the "true" sentence WER. In this scenario, we perform the adaptation with variable amounts of data, which are characterised by different levels of quality. Then, we move to realistic conditions in which the manual transcriptions of the evaluation data are not available. In this case, the adaptation is performed on data selected according to the WER scores "predicted" by a QE component. Our results indicate that: i) QE predictions allow us to closely approximate the adaptation results obtained in oracle conditions, and ii) the overall ASR performance based on the proposed QE-driven adaptation method is significantly better than the strong, most recent, CHiME-3 baseline.