[PDF][PDF] Keyword Spotting with Synthetic Data using Heterogeneous Knowledge Distillation.
Y Lee, SJ Baek - INTERSPEECH, 2022 - isca-archive.org
Y Lee, SJ Baek
INTERSPEECH, 2022•isca-archive.orgIt is crucial that Keyword Spotting (KWS) systems learn to understand new classes of user-
defined keywords, which however is a challenging task requiring high-quality audio
datasets. We propose KWS with Heterogeneous Embedding Knowledge Distillation (HEKD)
which uses only synthetic data of unseen
defined keywords, which however is a challenging task requiring high-quality audio
datasets. We propose KWS with Heterogeneous Embedding Knowledge Distillation (HEKD)
which uses only synthetic data of unseen
Abstract
It is crucial that Keyword Spotting (KWS) systems learn to understand new classes of user-defined keywords, which however is a challenging task requiring high-quality audio datasets. We propose KWS with Heterogeneous Embedding Knowledge Distillation (HEKD) which uses only synthetic data of unseen
isca-archive.org
以上显示的是最相近的搜索结果。 查看全部搜索结果