Word Error Rate Estimation for Speech Recognition: e-WER

<p dir="ltr">Measuring the performance of automatic speech recognition (ASR) systems requires manually transcribed data in order to compute the word error rate (WER), which is often time-consuming and expensive. In this paper, we propose a novel approach to estimate WER, or e-WER, wh...

وصف كامل

محفوظ في:
التفاصيل البيبلوغرافية
المؤلف الرئيسي: Ahmed Ali (705582) (author)
مؤلفون آخرون: Steve Renals (17196391) (author)
منشور في: 2018
الموضوعات:
الوسوم: إضافة وسم
لا توجد وسوم, كن أول من يضع وسما على هذه التسجيلة!
الوصف
الملخص:<p dir="ltr">Measuring the performance of automatic speech recognition (ASR) systems requires manually transcribed data in order to compute the word error rate (WER), which is often time-consuming and expensive. In this paper, we propose a novel approach to estimate WER, or e-WER, which does not require a gold-standard transcription of the test set. Our e-WER framework uses a comprehensive set of features: ASR recognised text, character recognition results to complement recognition output, and internal decoder features. We report results for the two features; black-box and glass-box using unseen 24 Arabic broadcast programs. Our system achieves 16.9% WER root mean squared error (RMSE) across 1,400 sentences. The estimated overall WER eWER was 25.3% for the three hours test set, while the actual WER was 28.5%.</p><h2>Other Information</h2><p dir="ltr">Published in: Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)<br>License: <a href="https://creativecommons.org/licenses/by/4.0/" target="_blank">https://creativecommons.org/licenses/by/4.0/</a><br>See article on publisher's website: <a href="https://dx.doi.org/10.18653/v1/p18-2004" target="_blank">https://dx.doi.org/10.18653/v1/p18-2004</a></p>