[b]=[d]-[t]+[p]: Self-supervised Speech Models Discover Phonological Vector Arithmetic
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866911589062410240 |
|---|---|
| author | Choi, Kwanghee Yeo, Eunjung Cho, Cheol Jun Harwath, David Mortensen, David R. |
| author_facet | Choi, Kwanghee Yeo, Eunjung Cho, Cheol Jun Harwath, David Mortensen, David R. |
| contents | Self-supervised speech models (S3Ms) are known to encode rich phonetic information, yet how this information is structured remains underexplored. We conduct a comprehensive study across 96 languages to analyze the underlying structure of S3M representations, with particular attention to phonological vectors. We first show that there exist linear directions within the model's representation space that correspond to phonological features. We further demonstrate that the scale of these phonological vectors correlate to the degree of acoustic realization of their corresponding phonological features in a continuous manner. For example, the difference between [d] and [t] yields a voicing vector: adding this vector to [p] produces [b], while scaling it results in a continuum of voicing. Together, these findings indicate that S3Ms encode speech using phonologically interpretable and compositional vectors, demonstrating phonological vector arithmetic. All code and interactive demos are available at https://github.com/juice500ml/phonetic-arithmetic . |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_18899 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | [b]=[d]-[t]+[p]: Self-supervised Speech Models Discover Phonological Vector Arithmetic Choi, Kwanghee Yeo, Eunjung Cho, Cheol Jun Harwath, David Mortensen, David R. Audio and Speech Processing Computation and Language Machine Learning Sound Self-supervised speech models (S3Ms) are known to encode rich phonetic information, yet how this information is structured remains underexplored. We conduct a comprehensive study across 96 languages to analyze the underlying structure of S3M representations, with particular attention to phonological vectors. We first show that there exist linear directions within the model's representation space that correspond to phonological features. We further demonstrate that the scale of these phonological vectors correlate to the degree of acoustic realization of their corresponding phonological features in a continuous manner. For example, the difference between [d] and [t] yields a voicing vector: adding this vector to [p] produces [b], while scaling it results in a continuum of voicing. Together, these findings indicate that S3Ms encode speech using phonologically interpretable and compositional vectors, demonstrating phonological vector arithmetic. All code and interactive demos are available at https://github.com/juice500ml/phonetic-arithmetic . |
| title | [b]=[d]-[t]+[p]: Self-supervised Speech Models Discover Phonological Vector Arithmetic |
| topic | Audio and Speech Processing Computation and Language Machine Learning Sound |
| url | https://arxiv.org/abs/2602.18899 |