Research

Paper

TESTING March 06, 2026

ImKWS: Test-Time Adaptation for Keyword Spotting with Class Imbalance

Authors

Hanyu Ding, Yang Xiao, Jiaheng Dong, Ting Dang

Abstract

Keyword spotting (KWS) identifies words for voice assistants, but environmental noise frequently reduces accuracy. Standard adaptation fixes this issue and strictly requires original or labeled audio. Test time adaptation (TTA) solves this data constraint using only unlabeled test audio. However, current methods fail to handle the severe imbalance between rare keywords and frequent background sounds. Consequently, standard entropy minimization (EM) becomes overconfident and heavily biased toward the frequent background class. To overcome this problem, we propose a TTA method named ImKWS. Our approach splits the entropy process into a reward branch and a penalty branch with separate update strengths. Furthermore, we enforce consistency across multiple audio transformations to ensure stable model updates. Experiments on the Google Speech Commands dataset indicate ImKWS achieves reliable adaptation in realistic imbalanced scenarios. The code is available on GitHub.

Metadata

arXiv ID: 2603.05821
Provider: ARXIV
Primary Category: eess.AS
Published: 2026-03-06
Fetched: 2026-03-09 06:05

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.05821v1</id>\n    <title>ImKWS: Test-Time Adaptation for Keyword Spotting with Class Imbalance</title>\n    <updated>2026-03-06T02:08:41Z</updated>\n    <link href='https://arxiv.org/abs/2603.05821v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.05821v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>Keyword spotting (KWS) identifies words for voice assistants, but environmental noise frequently reduces accuracy. Standard adaptation fixes this issue and strictly requires original or labeled audio. Test time adaptation (TTA) solves this data constraint using only unlabeled test audio. However, current methods fail to handle the severe imbalance between rare keywords and frequent background sounds. Consequently, standard entropy minimization (EM) becomes overconfident and heavily biased toward the frequent background class. To overcome this problem, we propose a TTA method named ImKWS. Our approach splits the entropy process into a reward branch and a penalty branch with separate update strengths. Furthermore, we enforce consistency across multiple audio transformations to ensure stable model updates. Experiments on the Google Speech Commands dataset indicate ImKWS achieves reliable adaptation in realistic imbalanced scenarios. The code is available on GitHub.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='eess.AS'/>\n    <published>2026-03-06T02:08:41Z</published>\n    <arxiv:comment>Submitted to Interspeech</arxiv:comment>\n    <arxiv:primary_category term='eess.AS'/>\n    <author>\n      <name>Hanyu Ding</name>\n    </author>\n    <author>\n      <name>Yang Xiao</name>\n    </author>\n    <author>\n      <name>Jiaheng Dong</name>\n    </author>\n    <author>\n      <name>Ting Dang</name>\n    </author>\n  </entry>"
}