Generalized Prompt Tuning: Adapting Frozen Univariate Time Series Foundation Models for Multivariate Healthcare Time Series

Mingzhu Liu, Angela Chen, George Chen
Proceedings of the 4th Machine Learning for Health Symposium, PMLR 259:668-679, 2025.

Abstract

Time series foundation models are pre-trained on large datasets and are able to achieve state-of-the-art performance in diverse tasks. However, to date, there has been limited work demonstrating how well these models perform in medical applications, where labeled data can be scarce. Further, we observe that currently, the majority of time series foundation models either are univariate in nature, or assume channel independence, meaning that they handle multivariate time series but do not model how the different variables relate. In this paper, we propose a prompt-tuning-inspired fine-tuning technique, Generalized Prompt Tuning (Gen-P-Tuning), that enables us to adapt an existing univariate time series foundation model (treated as frozen) to handle multivariate time series prediction. Our approach provides a way to combine information across channels (variables) of multivariate time series. We demonstrate the effectiveness of our fine-tuning approach against various baselines on two MIMIC classification tasks, and on influenza-like illness forecasting.

Cite this Paper

BibTeX
@InProceedings{pmlr-v259-liu25a, title = {Generalized Prompt Tuning: Adapting Frozen Univariate Time Series Foundation Models for Multivariate Healthcare Time Series}, author = {Liu, Mingzhu and Chen, Angela and Chen, George}, booktitle = {Proceedings of the 4th Machine Learning for Health Symposium}, pages = {668--679}, year = {2025}, editor = {Hegselmann, Stefan and Zhou, Helen and Healey, Elizabeth and Chang, Trenton and Ellington, Caleb and Mhasawade, Vishwali and Tonekaboni, Sana and Argaw, Peniel and Zhang, Haoran}, volume = {259}, series = {Proceedings of Machine Learning Research}, month = {15--16 Dec}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v259/main/assets/liu25a/liu25a.pdf}, url = {https://proceedings.mlr.press/v259/liu25a.html}, abstract = {Time series foundation models are pre-trained on large datasets and are able to achieve state-of-the-art performance in diverse tasks. However, to date, there has been limited work demonstrating how well these models perform in medical applications, where labeled data can be scarce. Further, we observe that currently, the majority of time series foundation models either are univariate in nature, or assume channel independence, meaning that they handle multivariate time series but do not model how the different variables relate. In this paper, we propose a prompt-tuning-inspired fine-tuning technique, Generalized Prompt Tuning (Gen-P-Tuning), that enables us to adapt an existing univariate time series foundation model (treated as frozen) to handle multivariate time series prediction. Our approach provides a way to combine information across channels (variables) of multivariate time series. We demonstrate the effectiveness of our fine-tuning approach against various baselines on two MIMIC classification tasks, and on influenza-like illness forecasting.} }
Endnote
%0 Conference Paper %T Generalized Prompt Tuning: Adapting Frozen Univariate Time Series Foundation Models for Multivariate Healthcare Time Series %A Mingzhu Liu %A Angela Chen %A George Chen %B Proceedings of the 4th Machine Learning for Health Symposium %C Proceedings of Machine Learning Research %D 2025 %E Stefan Hegselmann %E Helen Zhou %E Elizabeth Healey %E Trenton Chang %E Caleb Ellington %E Vishwali Mhasawade %E Sana Tonekaboni %E Peniel Argaw %E Haoran Zhang %F pmlr-v259-liu25a %I PMLR %P 668--679 %U https://proceedings.mlr.press/v259/liu25a.html %V 259 %X Time series foundation models are pre-trained on large datasets and are able to achieve state-of-the-art performance in diverse tasks. However, to date, there has been limited work demonstrating how well these models perform in medical applications, where labeled data can be scarce. Further, we observe that currently, the majority of time series foundation models either are univariate in nature, or assume channel independence, meaning that they handle multivariate time series but do not model how the different variables relate. In this paper, we propose a prompt-tuning-inspired fine-tuning technique, Generalized Prompt Tuning (Gen-P-Tuning), that enables us to adapt an existing univariate time series foundation model (treated as frozen) to handle multivariate time series prediction. Our approach provides a way to combine information across channels (variables) of multivariate time series. We demonstrate the effectiveness of our fine-tuning approach against various baselines on two MIMIC classification tasks, and on influenza-like illness forecasting.
APA
Liu, M., Chen, A. & Chen, G.. (2025). Generalized Prompt Tuning: Adapting Frozen Univariate Time Series Foundation Models for Multivariate Healthcare Time Series. Proceedings of the 4th Machine Learning for Health Symposium, in Proceedings of Machine Learning Research 259:668-679 Available from https://proceedings.mlr.press/v259/liu25a.html.

Related Material