Cross-device Zero-shot Label Transfer via Alignment of Time Series Foundation Model Embeddings
Fuente:
arXiv
Salvato in:
| Autori principali: | , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866908525304741888 |
|---|---|
| author | Ravindra, Neal G. Sehanobish, Arijit |
| author_facet | Ravindra, Neal G. Sehanobish, Arijit |
| contents | High-quality, medically validated labels exist for clinical actigraphy data but not for ubiquitous consumer wearables like the Apple Watch. Manually labeling wearables data is expensive and doesn't scale. This paper offers a novel framework that transfers valuable labels from a source domain (e.g., actigraphy) to a target domain (e.g., Apple Watch) without requiring paired data. Instead of working with raw time-series signals, we project both domains into a shared latent embedding space using time-series foundation models (TSFMs) and develop a new framework to align the cross-device representations. Our method, Adversarial Alignment of TSFM Embeddings forces the distributions of source and target embeddings to align within this space, facilitating label transfer across device type. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_06966 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Cross-device Zero-shot Label Transfer via Alignment of Time Series Foundation Model Embeddings Ravindra, Neal G. Sehanobish, Arijit Signal Processing Artificial Intelligence Machine Learning High-quality, medically validated labels exist for clinical actigraphy data but not for ubiquitous consumer wearables like the Apple Watch. Manually labeling wearables data is expensive and doesn't scale. This paper offers a novel framework that transfers valuable labels from a source domain (e.g., actigraphy) to a target domain (e.g., Apple Watch) without requiring paired data. Instead of working with raw time-series signals, we project both domains into a shared latent embedding space using time-series foundation models (TSFMs) and develop a new framework to align the cross-device representations. Our method, Adversarial Alignment of TSFM Embeddings forces the distributions of source and target embeddings to align within this space, facilitating label transfer across device type. |
| title | Cross-device Zero-shot Label Transfer via Alignment of Time Series Foundation Model Embeddings |
| topic | Signal Processing Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2509.06966 |