Few-Shot Open-Set Learning for On-Device Customization of KeyWord Spotting Systems
This work addresses the challenge of personalizing keyword spotting for users without requiring large datasets, though it is incremental as it builds on existing few-shot and prototype-based methods.
The paper tackles the problem of enabling fast on-device customization for keyword spotting systems by using few-shot learning to classify user-defined keywords with limited data, achieving up to 76% accuracy in a 10-shot scenario while maintaining a 5% false acceptance rate for unknown data.
A personalized KeyWord Spotting (KWS) pipeline typically requires the training of a Deep Learning model on a large set of user-defined speech utterances, preventing fast customization directly applied on-device. To fill this gap, this paper investigates few-shot learning methods for open-set KWS classification by combining a deep feature encoder with a prototype-based classifier. With user-defined keywords from 10 classes of the Google Speech Command dataset, our study reports an accuracy of up to 76% in a 10-shot scenario while the false acceptance rate of unknown data is kept to 5%. In the analyzed settings, the usage of the triplet loss to train an encoder with normalized output features performs better than the prototypical networks jointly trained with a generator of dummy unknown-class prototypes. This design is also more effective than encoders trained on a classification problem and features fewer parameters than other iso-accuracy approaches.