arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2106.01604cs.LGstat.ML
收藏

嘈杂的学生-教师培训,用于强大的关键字识别

Noisy student-teacher training for robust keyword spotting

Google Inc.(谷歌公司)

Hyun-Jin Park, Pai Zhu, Ignacio Lopez Moreno, Niranjan Subrahmanya

摘要:

我们提出了自训练与嘈杂的学生-教师的方法流关键字发现,可以利用大规模的未标记的数据和积极的数据增强。所提出的方法应用积极的数据增强(光谱增强)的学生和教师的输入,并利用未标记的数据在规模,这显着提高了学生的准确性,对具有挑战性的条件。当与带有硬标记数据的监督训练一起使用时,这种积极的增强通常会降低模型性能。实验表明,积极规范增强基线监督训练方法降低准确性,而提出的自我训练与嘈杂的学生-教师培训提高了一些困难的条件测试集的准确性高达60%。

英文摘要:

We propose self-training with noisy student-teacher approach for streaming keyword spotting, that can utilize large-scale unlabeled data and aggressive data augmentation. The proposed method applies aggressive data augmentation (spectral augmentation) on the input of both student and teacher and utilize unlabeled data at scale, which significantly boosts the accuracy of student against challenging conditions. Such aggressive augmentation usually degrades model performance when used with supervised training with hard-labeled data. Experiments show that aggressive spec augmentation on baseline supervised training method degrades accuracy, while the proposed self-training with noisy student-teacher training improves accuracy of some difficult-conditioned test sets by as much as 60%.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑