Does Whisper understand Swiss German? An automatic, qualitative, and human evaluation
This work addresses the need for reliable ASR in Swiss German, a low-resource dialect, but is incremental as it primarily evaluates an existing model without introducing new methods.
The paper tackled the problem of evaluating Whisper's automatic speech recognition performance on Swiss German dialects, finding that it transcribes Swiss German into Standard German with viable accuracy across multiple test sets, including a new mock clinical interview dataset.
Whisper is a state-of-the-art automatic speech recognition (ASR) model (Radford et al., 2022). Although Swiss German dialects are allegedly not part of Whisper's training data, preliminary experiments showed that Whisper can transcribe Swiss German quite well, with the output being a speech translation into Standard German. To gain a better understanding of Whisper's performance on Swiss German, we systematically evaluate it using automatic, qualitative, and human evaluation. We test its performance on three existing test sets: SwissDial (Dogan-Schönberger et al., 2021), STT4SG-350 (Plüss et al., 2023), and Swiss Parliaments Corpus (Plüss et al., 2021). In addition, we create a new test set for this work, based on short mock clinical interviews. For automatic evaluation, we used word error rate (WER) and BLEU. In the qualitative analysis, we discuss Whisper's strengths and weaknesses and anylyze some output examples. For the human evaluation, we conducted a survey with 28 participants who were asked to evaluate Whisper's performance. All of our evaluations suggest that Whisper is a viable ASR system for Swiss German, so long as the Standard German output is desired.