WERd: Using Social Text Spelling Variants for Evaluating Dialectal Speech Recognition
This addresses the challenge of evaluating ASR for dialectal speech, which is important for improving speech technology in linguistically diverse communities, but is incremental as it adapts an existing MT metric.
The authors tackled the problem of evaluating automatic speech recognition (ASR) systems for dialectal speech, where standard word error rate (WER) is inadequate due to non-standardized orthography, and developed WERd, a variant of TERp that uses unsupervised mining of spelling variants from Twitter, showing it is more adequate than WER for evaluating Egyptian Arabic ASR.
We study the problem of evaluating automatic speech recognition (ASR) systems that target dialectal speech input. A major challenge in this case is that the orthography of dialects is typically not standardized. From an ASR evaluation perspective, this means that there is no clear gold standard for the expected output, and several possible outputs could be considered correct according to different human annotators, which makes standard word error rate (WER) inadequate as an evaluation metric. Such a situation is typical for machine translation (MT), and thus we borrow ideas from an MT evaluation metric, namely TERp, an extension of translation error rate which is closely-related to WER. In particular, in the process of comparing a hypothesis to a reference, we make use of spelling variants for words and phrases, which we mine from Twitter in an unsupervised fashion. Our experiments with evaluating ASR output for Egyptian Arabic, and further manual analysis, show that the resulting WERd (i.e., WER for dialects) metric, a variant of TERp, is more adequate than WER for evaluating dialectal ASR.