Can I read Representation Learning for Keyword Spotting on EtoBox?

Representation Learning for Keyword Spotting by Valentina Fedchenko is a document available to read on EtoBox.

What is Representation Learning for Keyword Spotting about?

This paper explores representation learning techniques for low-resource keyword spotting (KWS), addressing challenges such as limited labeled data and device resources. The authors propose a local-global contrastive siamese network (LGCSiam) and a self-supervised Wav2Vec 2.0 model to enhance the performance of small-footprint KWS models through self-training and representation learning. Experimental results demonstrate significant improvements in accuracy, particularly when training on small labeled dataset

Author

Valentina Fedchenko

Language

EN