Language in the Flow of Time: Time-Series-Paired Texts Weaved into a Unified Temporal Narrative
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_ | 1866917319767228416 |
|---|---|
| author | Li, Zihao Lin, Xiao Liu, Zhining Zou, Jiaru Wu, Ziwei Zheng, Lecheng Fu, Dongqi Zhu, Yada Hamann, Hendrik Tong, Hanghang He, Jingrui |
| author_facet | Li, Zihao Lin, Xiao Liu, Zhining Zou, Jiaru Wu, Ziwei Zheng, Lecheng Fu, Dongqi Zhu, Yada Hamann, Hendrik Tong, Hanghang He, Jingrui |
| contents | While many advances in time series models focus exclusively on numerical data, research on multimodal time series, particularly those involving contextual textual information, remains in its infancy. With recent progress in large language models and time series learning, we revisit the integration of paired texts with time series through the Platonic Representation Hypothesis, which posits that representations of different modalities converge to shared spaces. In this context, we identify that time-series-paired texts may naturally exhibit periodic properties that closely mirror those of the original time series. Building on this insight, we propose a novel framework, Texts as Time Series (TaTS), which considers the time-series-paired texts to be auxiliary variables of the time series. TaTS can be plugged into any existing numerical-only time series models and effectively enable them to handle time series data with paired texts. Through extensive experiments on both multimodal time series forecasting and imputation tasks across benchmark datasets with various existing time series models, we demonstrate that TaTS can enhance multimodal predictive performance without modifying model architectures. Our Code is available at https://github.com/iDEA-iSAIL-Lab-UIUC/TaTS. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2502_08942 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Language in the Flow of Time: Time-Series-Paired Texts Weaved into a Unified Temporal Narrative Li, Zihao Lin, Xiao Liu, Zhining Zou, Jiaru Wu, Ziwei Zheng, Lecheng Fu, Dongqi Zhu, Yada Hamann, Hendrik Tong, Hanghang He, Jingrui Machine Learning Artificial Intelligence While many advances in time series models focus exclusively on numerical data, research on multimodal time series, particularly those involving contextual textual information, remains in its infancy. With recent progress in large language models and time series learning, we revisit the integration of paired texts with time series through the Platonic Representation Hypothesis, which posits that representations of different modalities converge to shared spaces. In this context, we identify that time-series-paired texts may naturally exhibit periodic properties that closely mirror those of the original time series. Building on this insight, we propose a novel framework, Texts as Time Series (TaTS), which considers the time-series-paired texts to be auxiliary variables of the time series. TaTS can be plugged into any existing numerical-only time series models and effectively enable them to handle time series data with paired texts. Through extensive experiments on both multimodal time series forecasting and imputation tasks across benchmark datasets with various existing time series models, we demonstrate that TaTS can enhance multimodal predictive performance without modifying model architectures. Our Code is available at https://github.com/iDEA-iSAIL-Lab-UIUC/TaTS. |
| title | Language in the Flow of Time: Time-Series-Paired Texts Weaved into a Unified Temporal Narrative |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2502.08942 |