Keyword spotting accuracy degrades when neural networks are exposed to noisy
environments. On-site adaptation to previously unseen noise is crucial to
recovering accuracy loss, and on-device learning is required to ensure that the
adaptation process happens entirely on the edge device. In this work, we
propose a fully on-device domain adaptation system achieving up to 14% accuracy
gains over already-robust keyword spotting models. We enable on-device learning
with less than 10 kB of memory, using only 100 labeled utterances to recover 5%
accuracy after adapting to the complex speech noise. We demonstrate that domain
adaptation can be achieved on ultra-low-power microcontrollers with as little
as 806 mJ in only 14 s on always-on, battery-operated devices.

在本研究中，我们提出了一个完全基于设备的领域适应系统，相较于已经稳健的关键词识别模型，能够获得 14% 的准确度提升。我们使用少于 10 kB 的内存，仅 100 个标记的话语就能在适应复杂的语音噪声后恢复 5% 的准确度，展示了域适应在低功耗微控制器上的可行性，仅需 806 mJ 的能量，在始终开启的电池供电设备上执行时间为 14 秒。