|کد مقاله||کد نشریه||سال انتشار||مقاله انگلیسی||ترجمه فارسی||نسخه تمام متن|
|4973647||1451680||2018||26 صفحه PDF||سفارش دهید||دانلود کنید|
- Review on automatic speech recognition quality estimation (ASR QE).
- The application of ASR QE in ASR system combination for both single-microphone multiple-ASR system task and on multiple-microphone multiple-ASR system task.
- Ranking the system combination inputs based on predicted quality.
- Management of tied ranks.
- Automatically finding the optimum level of combination for each segment.
Recognizer Output Voting Error Reduction (ROVER) has been widely used for system combination in automatic speech recognition (ASR). In order to select the most appropriate words to insert at each position in the output transcriptions, some ROVER extensions rely on critical information such as confidence scores and other ASR decoder features. This information, which is not always available, highly depends on the decoding process and sometimes tends to overestimate the real quality of the recognized words. In this paper we propose a novel variant of ROVER that takes advantage of ASR quality estimation (QE) for ranking the transcriptions at “segment level” instead of: i) relying on confidence scores, or ii) feeding ROVER with randomly ordered hypotheses. We first introduce an effective set of features to compensate for the absence of ASR decoder information. Then, we apply QE techniques to perform accurate hypothesis ranking at segment-level before starting the fusion process. The evaluation is carried out on two different tasks, in which we respectively combine hypotheses coming from independent ASR systems and multi-microphone recordings. In both tasks, it is assumed that the ASR decoder information is not available. The proposed approach significantly outperforms standard ROVER and it is competitive with two strong oracles that exploit prior knowledge about the real quality of the hypotheses to be combined. Compared to standard ROVER, the absolute WER improvements in the two evaluation scenarios range from 0.5% to 7.3%.
Journal: Computer Speech & Language - Volume 47, January 2018, Pages 214-239