Mitigating Accent-Language Confusion in Self-Supervised Speech Representations for Language Identification

Key points

• Spoken language identification (LID) aims to recognize the target language regardless of accent.1
• In practice, however, LID models fine-tuned from self-supervised speech representations frequently confuse accents with languages, misclassifying non-native (L2) speech as the speaker's first language (L1).1
• To address this, we introduce a geometric projection that estimates an L1-bias direction solely from native speech and removes it before the frozen LID head.1
• Across five MMS-LID models and non-native corpora, this projection substantially improves target language identification for L2-accented speech while preserving predictions for native speech.1

Sources (1)

• [1]Mitigating Accent-Language Confusion in Self-Supervised Speech Representations for Language Identification
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 7, 05:42 AM

  We show that non-native speech representations lie between native target-language and native L1 poles, causing systematic misclassification.

  Spoken language identification (LID) aims to recognize the target language regardless of accent.

Extractive summary: sentences quoted from the sources.