Research

Paper

TESTING March 09, 2026

Quantifying Cross-Lingual Transfer in Paralinguistic Speech Tasks

Authors

Pol Buitrago, Oriol Pareras, Federico Costa, Javier Hernando

Abstract

Paralinguistic speech tasks are often considered relatively language-agnostic, as they rely on extralinguistic acoustic cues rather than lexical content. However, prior studies report performance degradation under cross-lingual conditions, indicating non-negligible language dependence. Still, these studies typically focus on isolated language pairs or task-specific settings, limiting comparability and preventing a systematic assessment of task-level language dependence. We introduce the Cross-Lingual Transfer Matrix (CLTM), a systematic method to quantify cross-lingual interactions between pairs of languages within a given task. We apply the CLTM to two paralinguistic tasks, gender identification and speaker verification, using a multilingual HuBERT-based encoder, to analyze how donor-language data affects target-language performance during fine-tuning. Our results reveal distinct transfer patterns across tasks and languages, reflecting systematic, language-dependent effects.

Metadata

arXiv ID: 2603.08231
Provider: ARXIV
Primary Category: eess.AS
Published: 2026-03-09
Fetched: 2026-03-10 05:43

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.08231v1</id>\n    <title>Quantifying Cross-Lingual Transfer in Paralinguistic Speech Tasks</title>\n    <updated>2026-03-09T11:02:57Z</updated>\n    <link href='https://arxiv.org/abs/2603.08231v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.08231v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>Paralinguistic speech tasks are often considered relatively language-agnostic, as they rely on extralinguistic acoustic cues rather than lexical content. However, prior studies report performance degradation under cross-lingual conditions, indicating non-negligible language dependence. Still, these studies typically focus on isolated language pairs or task-specific settings, limiting comparability and preventing a systematic assessment of task-level language dependence.\n  We introduce the Cross-Lingual Transfer Matrix (CLTM), a systematic method to quantify cross-lingual interactions between pairs of languages within a given task. We apply the CLTM to two paralinguistic tasks, gender identification and speaker verification, using a multilingual HuBERT-based encoder, to analyze how donor-language data affects target-language performance during fine-tuning. Our results reveal distinct transfer patterns across tasks and languages, reflecting systematic, language-dependent effects.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='eess.AS'/>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.CL'/>\n    <published>2026-03-09T11:02:57Z</published>\n    <arxiv:comment>6 pages, 5 figures, Submitted to Interspeech 2026</arxiv:comment>\n    <arxiv:primary_category term='eess.AS'/>\n    <author>\n      <name>Pol Buitrago</name>\n    </author>\n    <author>\n      <name>Oriol Pareras</name>\n    </author>\n    <author>\n      <name>Federico Costa</name>\n    </author>\n    <author>\n      <name>Javier Hernando</name>\n    </author>\n  </entry>"
}