From Weak Labels to Strong Results: Utilizing 5,000 Hours of Noisy Classroom Transcripts with Minimal Accurate Data
This addresses the challenge of high transcription costs in classroom ASR, offering a practical solution for low-resource settings, though it is incremental as it builds on existing pretraining and fine-tuning paradigms.
The paper tackled the problem of low-resource classroom Automatic Speech Recognition (ASR) by proposing Weakly Supervised Pretraining (WSP), a two-step method that first pretrains on abundant noisy transcripts and then fine-tunes on limited accurate data, resulting in improved performance over alternative methods.
Recent progress in speech recognition has relied on models trained on vast amounts of labeled data. However, classroom Automatic Speech Recognition (ASR) faces the real-world challenge of abundant weak transcripts paired with only a small amount of accurate, gold-standard data. In such low-resource settings, high transcription costs make re-transcription impractical. To address this, we ask: what is the best approach when abundant inexpensive weak transcripts coexist with limited gold-standard data, as is the case for classroom speech data? We propose Weakly Supervised Pretraining (WSP), a two-step process where models are first pretrained on weak transcripts in a supervised manner, and then fine-tuned on accurate data. Our results, based on both synthetic and real weak transcripts, show that WSP outperforms alternative methods, establishing it as an effective training methodology for low-resource ASR in real-world scenarios.