Training Wake Word Detection with Synthesized Speech Data on Confusion Words

TL;DR

This paper explores data augmentation techniques, including synthesized speech and noise addition, to improve wake word detection robustness against confusing words, demonstrating significant performance gains in challenging scenarios.

Contribution

It introduces the use of multi-speaker synthesized speech data for training wake word detection systems, enhancing their robustness against confusing words.

Findings

Synthetic data improves detection accuracy on confusing words

Adding noise to features enhances system robustness

Synthetic augmentation yields significant performance gains

Abstract

Confusing-words are commonly encountered in real-life keyword spotting applications, which causes severe degradation of performance due to complex spoken terms and various kinds of words that sound similar to the predefined keywords. To enhance the wake word detection system's robustness on such scenarios, we investigate two data augmentation setups for training end-to-end KWS systems. One is involving the synthesized data from a multi-speaker speech synthesis system, and the other augmentation is performed by adding random noise to the acoustic feature. Experimental results show that augmentations help improve the system's robustness. Moreover, by augmenting the training set with the synthetic data generated by the multi-speaker text-to-speech system, we achieve a significant improvement regarding confusing words scenario.