Salvato in:
| Autori principali: | , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2511.12285 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866911395202727936 |
|---|---|
| author | Kim, Minu Um, Ji Sub Kim, Hoirin |
| author_facet | Kim, Minu Um, Ji Sub Kim, Hoirin |
| contents | Lexical tone is central to many languages but remains underexplored in self-supervised learning (SSL) speech models, especially beyond Mandarin. We study four languages with complex and diverse tone systems (Burmese, Thai, Lao, and Vietnamese) to ask how far such models "listen" for tone and how transfer operates in low-resource conditions. As a baseline reference, we estimate the temporal span of tone cues: approximately 100ms (Burmese/Thai) and 180ms (Lao/Vietnamese). Probes and gradient analysis on fine-tuned SSL models reveal that tone transfer varies by downstream task: automatic speech recognition fine-tuning aligns spans with language-specific tone cues, while prosody- and voice-related tasks bias toward overly long spans. These findings indicate that tone transfer is shaped by downstream task, highlighting task effects on temporal focus in tone modeling. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_12285 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | How Far Do SSL Speech Models Listen for Tone? Temporal Focus of Tone Representation under Low-resource Transfer Kim, Minu Um, Ji Sub Kim, Hoirin Audio and Speech Processing Computation and Language Lexical tone is central to many languages but remains underexplored in self-supervised learning (SSL) speech models, especially beyond Mandarin. We study four languages with complex and diverse tone systems (Burmese, Thai, Lao, and Vietnamese) to ask how far such models "listen" for tone and how transfer operates in low-resource conditions. As a baseline reference, we estimate the temporal span of tone cues: approximately 100ms (Burmese/Thai) and 180ms (Lao/Vietnamese). Probes and gradient analysis on fine-tuned SSL models reveal that tone transfer varies by downstream task: automatic speech recognition fine-tuning aligns spans with language-specific tone cues, while prosody- and voice-related tasks bias toward overly long spans. These findings indicate that tone transfer is shaped by downstream task, highlighting task effects on temporal focus in tone modeling. |
| title | How Far Do SSL Speech Models Listen for Tone? Temporal Focus of Tone Representation under Low-resource Transfer |
| topic | Audio and Speech Processing Computation and Language |
| url | https://arxiv.org/abs/2511.12285 |