Learning from Time Series for Health

Time series data are ubiquitous in healthcare, from medical time series to wearable data, and present an exciting opportunity for machine learning methods to extract actionable insights about human health. However, huge gap remain between the existing time series literature and what is needed to make machine learning systems practical and deployable for healthcare. This is because learning from time series for health is notoriously challenging: labels are often noisy or missing, data can be multimodal and extremely high dimensional, missing values are pervasive, measurements are irregular, data distributions shift rapidly over time, explaining model outcomes is challenging, and deployed models require careful maintenance over time. These challenges introduce interesting research problems that the community has been actively working on for the last few years, with significant room for contribution still remaining. Learning from time series for health is a uniquely challenging and important area with increasing application. Significant advancements are required to realize the societal benefits of these systems for healthcare. This workshop will bring together machine learning researchers dedicated to advancing the field of time series modeling in healthcare to bring these models closer to deployment.

Call for Papers

In our Time Series for Health Workshop, we delve into the complexities of time series data to better understand and improve human health. This field boasts rich diversity, encompassing various modalities such as wearables, Electronic Health Record (EHR) data, medical time series including ECG, EEG, fMRI, and audio data. Our workshop will pivot around two central themes: Behavioral Health: Exploring the intricate dynamics of behavioral patterns and their implications on health through time series analysis. Foundation Models: Investigating the core models that form the bedrock for understanding and interpreting time series data in healthcare. These themes will be echoed in our keynote addresses, round-tables, and interactive panel discussions. Submissions that align with these themes will be given special consideration for spotlight talks. However, all submissions that meet the guidelines listed below will be considered.

Submission Guidelines We invite papers that:

Propose innovative methods or perspectives.
Present preliminary results that open avenues for future research.
Introduce new resources like datasets to propel research in this domain.
Clearly demonstrate or discuss their relevance to healthcare, specifically focusing on challenges within health time series data.

Topics of Interest Submissions may address, but are not limited to the follow topics as they relate to time series:

Unsupervised, semi-supervised, and supervised representation learning.
Novel architectures or models.
Classification, regression, and forecasting.
Bayesian models.
Sequential decision-making.
Challenges of time series data: missing values, noisy/irregular measurements, high-dimensionality.
Multi-modal models incorporating time series.
Deployment and implementation challenges.
Explainability, fairness, and privacy in time series models.
Practical applications (e.g., dynamic treatment recommendation for sepsis from EHR time series).

Ethical Considerations We encourage discussions on ethical aspects pertinent to your research. Submission Instructions We invite extended abstract submissions that are up to 4 pages long (not including references). All accepted papers will be presented in person as posters. There are no formal proceedings generated from this workshop, though all papers will be made public upon their acceptance. Authors are encouraged to make their work publicly available through our online listing of presented work. The reviewing process will be double-blind. Please submit anonymized versions of your paper that include no identifying information about any author identities or affiliations. Submitted papers must be new work that has not yet been published. We allow submissions that are already on arXiv and reviewers will be asked to not seek out public versions of papers during the reviewing process.

Submission format: ICLR 2024 paper style

Submission link: OpenReview

Important Dates