An End-to-End Architecture for Keyword Spotting and Voice Activity Detection

TL;DR

This paper introduces a unified neural network architecture that performs keyword spotting and voice activity detection simultaneously, achieving high accuracy without needing aligned training data or separate models.

Contribution

The authors present a novel end-to-end RNN model with inference algorithms that handle both tasks with shared parameters, eliminating the need for retraining or additional resources.

Findings

High accuracy on both tasks achieved

No requirement for aligned training data

Shared parameters enable efficient deployment

Abstract

We propose a single neural network architecture for two tasks: on-line keyword spotting and voice activity detection. We develop novel inference algorithms for an end-to-end Recurrent Neural Network trained with the Connectionist Temporal Classification loss function which allow our model to achieve high accuracy on both keyword spotting and voice activity detection without retraining. In contrast to prior voice activity detection models, our architecture does not require aligned training data and uses the same parameters as the keyword spotting model. This allows us to deploy a high quality voice activity detector with no additional memory or maintenance requirements.