Training Keyword Spotting Models on Non-IID Data with Federated Learning
This addresses the problem of on-device model training with privacy constraints for keyword spotting applications, but it is incremental as it builds on existing federated learning and optimization techniques.
The paper tackled training keyword spotting models on non-IID data using federated learning, achieving comparable false accept and false reject rates to centrally-trained models and reducing false reject rate by 56% with SpecAugment.
We demonstrate that a production-quality keyword-spotting model can be trained on-device using federated learning and achieve comparable false accept and false reject rates to a centrally-trained model. To overcome the algorithmic constraints associated with fitting on-device data (which are inherently non-independent and identically distributed), we conduct thorough empirical studies of optimization algorithms and hyperparameter configurations using large-scale federated simulations. To overcome resource constraints, we replace memory intensive MTR data augmentation with SpecAugment, which reduces the false reject rate by 56%. Finally, to label examples (given the zero visibility into on-device data), we explore teacher-student training.