Human-like cross-language generalisation in deep neural speaker embeddings and its acoustic foundations
Paper · https://arxiv.org/abs/2611.05384
arXiv:2611.05384v1 Announce Type: new Abstract: Generalisability is a central question for deep neural networks, including the cross-language generalisation of speaker embeddings. This study examined whether deep neural speaker embeddings generalise across languages and how such generalisation relates to human perception and acoustic structure. Using Cantonese--English bilingual speech presented in within- and cross-language conditions, we derived similarity from 18 state-of-the-art models and compared it with human ratings and acoustic features using linear mixed models and representational similarity analysis. Embeddings exhibited cross-language generalisation that broadly paralleled human behaviour, reflecting accurate identity discrimination and sensitivity to language switching. This generalisation relied on largely shared acoustic cues with language-sensitive weighting, mirroring human patterns. Together, these findings provide converging evidence for human--machine alignment in cross-language voice similarity processing.
First seen by Oriel on .
AI summary
The study finds that deep neural speaker embeddings generalise across Cantonese and English in ways that broadly parallel human voice-similarity judgments.
- Researchers evaluated similarity from 18 state-of-the-art models on bilingual speech in within-language and cross-language conditions.
- Embedding similarities were compared with human ratings and acoustic features using linear mixed models and representational similarity analysis.
- Models accurately discriminated speaker identity while remaining sensitive to language switching.
- Cross-language generalisation relied on largely shared acoustic cues, weighted differently depending on language, as in human judgments.
#speaker embeddings #speech processing #cross-language generalisation
Generated by gpt-5.6-sol. Check the original for details.
Aliases
| Source | External ID | URL |
|---|---|---|
| arXiv AI Papers | oai:arXiv.org:2611.05384v1 | https://arxiv.org/abs/2611.05384 Read discussion |