Mi-Go: Test Framework which uses YouTube as Data Source for Evaluating Speech Recognition Models like OpenAI's Whisper
This provides a practical tool for developers and researchers to assess speech recognition models across varied conditions, though it is incremental as it applies an existing method to new data.
The researchers tackled the problem of evaluating speech recognition models in diverse real-world scenarios by introducing Mi-Go, a testing framework that uses YouTube videos as a data source, and demonstrated it on OpenAI's Whisper models with 124 videos, showing YouTube's utility for robustness and accuracy testing.
This article introduces Mi-Go, a novel testing framework aimed at evaluating the performance and adaptability of general-purpose speech recognition machine learning models across diverse real-world scenarios. The framework leverages YouTube as a rich and continuously updated data source, accounting for multiple languages, accents, dialects, speaking styles, and audio quality levels. To demonstrate the effectiveness of the framework, the Whisper model, developed by OpenAI, was employed as a test object. The tests involve using a total of 124 YouTube videos to test all Whisper model versions. The results underscore the utility of YouTube as a valuable testing platform for speech recognition models, ensuring their robustness, accuracy, and adaptability to diverse languages and acoustic conditions. Additionally, by contrasting the machine-generated transcriptions against human-made subtitles, the Mi-Go framework can help pinpoint potential misuse of YouTube subtitles, like Search Engine Optimization.