Can I read Production Federated Keyword Spotting via Distillation, Filtering, and Joint Federated-centralized Training on EtoBox?

Production Federated Keyword Spotting via Distillation, Filtering, and Joint Federated-centralized Training by Hard, Andrew; Partridge, Kurt; Chen, Neng; Augenstein, Sean; Shah, Aishanee; Park, Hyun Jin; Park, Alex; Ng, Sara; Nguyen, Jessica; Moreno, Ignacio Lopez; Mathews, Rajiv; Beaufays, Françoise is a scholarly article available to read on EtoBox.

What is Production Federated Keyword Spotting via Distillation, Filtering, and Joint Federated-centralized Training about?

We trained a keyword spotting model using federated learning on real user devices and observed significant improvements when the model was deployed for inference on phones. To compensate for data domains that are missing from on-device training caches, we employed joint federated-centralized training. And to learn in the absence of curated labels on-device, we formulated a confidence filtering strategy based on user-feedback signals for federated distillation. These techniques created models that significantly improved quality metrics in offline evaluations and user-experience metrics in live A/B experiments.

Author

Hard, Andrew; Partridge, Kurt; Chen, Neng; Augenstein, Sean; Shah, Aishanee; Park, Hyun Jin; Park, Alex; Ng, Sara; Nguyen, Jessica; Moreno, Ignacio Lopez; Mathews, Rajiv; Beaufays, Françoise

Published

2022

Language

EN