GE2E-KWS: Generalized End-to-End Training and Evaluation for Zero-shot Keyword Spotting
This work addresses the need for efficient, on-device keyword spotting that can handle new keywords without retraining, though it is incremental as it builds on existing end-to-end and loss function approaches.
The authors tackled the problem of zero-shot keyword spotting by proposing GE2E-KWS, a generalized end-to-end training and evaluation framework, which resulted in a 419KB quantized conformer model outperforming a 7.5GB ASR encoder by 23.6% relative AUC and a same-size triplet loss model by 60.7% AUC.
We propose GE2E-KWS -- a generalized end-to-end training and evaluation framework for customized keyword spotting. Specifically, enrollment utterances are separated and grouped by keywords from the training batch and their embedding centroids are compared to all other test utterance embeddings to compute the loss. This simulates runtime enrollment and verification stages, and improves convergence stability and training speed by optimizing matrix operations compared to SOTA triplet loss approaches. To benchmark different models reliably, we propose an evaluation process that mimics the production environment and compute metrics that directly measure keyword matching accuracy. Trained with GE2E loss, our 419KB quantized conformer model beats a 7.5GB ASR encoder by 23.6% relative AUC, and beats a same size triplet loss model by 60.7% AUC. Our KWS models are natively streamable with low memory footprints, and designed to continuously run on-device with no retraining needed for new keywords (zero-shot).