The 1st International Workshop-Tutorial on Learning on Real and Synthetic Medical Time Series Data (MED-TIME) will be colocated with ECML-PKDD 2025 in Porto, Portugal.
This workshop will investigate the challenges and advances in machine learning for medical time series data, which emerge in patient monitoring, disease progression modeling, and clinical decision support for patients with chronic conditions. A key challenge in working with real medical data is handling sparsity and missingness, as the more people are recorded, the less data is there for each one. While synthetic data generation seems like a solution, synthetic medical time series pose challenges themselves, as ensuring its utility alongside real data requires careful validation. At the same time, multimodal data integration, explainability, and forecasting remain critical for developing reliable and interpretable models. This workshop will bring together experts working on time series, medical timestamped data, and the generation of synthetic data, temporal and static, to discuss research and open challenges in the field. The main focus of this workshop-tutorial will be on the following five themes:
- Learning on real and synthetic medical time series with emphasis on handling irregularity, sparsity, and missing values and on augmenting the few real data entities with many synthetic ones
- Challenges in synthetic medical time series generation including proximity to real data (and ways of measuring this), diversity, and privacy preservation
- Benchmarks for synthetic and real medical data and for combinations thereof
- Explainability in medical time series models, ensuring interpretability for medical experts
- Forecasting and integration of multiple modalities focusing on effective modeling across diverse medical data sources
Scope
The workshop will cover a broad range of topics at the intersection of data science and healthcare, including but not limited to:
- Methods for generating high-quality synthetic medical time series data
- Evaluation and benchmarking of synthetic medical data
- Machine learning approaches for real-world medical time series
- Data sparsity, missing values, and irregularity handling in medical time series
- Multimodal data fusion for improving medical time series analysis
- Privacy-preserving synthetic data for healthcare applications
- Explainability and interpretability in machine learning for medical time series
- Robustness and bias considerations in synthetic vs. real medical data
- Use cases and applications of synthetic medical data in clinical settings
Submissions will undergo a single-blind review process; author identities are known to reviewers. Each paper will be reviewed by at least three independent reviewers, while a senior PC member will monitor the reviewing process.
All accepted paper contributions will be published in the Lecture Notes in Computer Science series by Springer-Verlag
Formatting instructions
All submissions will undergo double-blind peer review. Accepted papers will be presented orally or as posters. Proceedings will be published in Springer’s Communications in Computer and Information Science (CCIS).
Contributions written in English must be formatted according to the guidelines of the Lecture Notes of Computer Science (LNCS) series by Springer-Verlag. These guidelines, along with templates, are available here: https://www.springer.com/gp/computer-science/lncs/conference-proceedings-guidelines.
Regular Papers (8-16 pages)
Extended Abstracts (2-4 pages)
Position Papers (2 pages) presenting novel ideas or challenges
Authors may not submit any paper that is under review elsewhere or that has been accepted for publication in a journal or another conference; neither will they submit their papers elsewhere during the review period of MED-TIME 2025. The authors, their contact information, or affiliations may not be changed after the paper submission deadline. If the paper is accepted, at least one of the authors must register for the Workshop and present the paper in person in Porto.
Submission
Papers must be submitted electronically in PDF format. The submission interface will be available here shortly.
- Paper submission deadline: June 14, 2025
- Notification of acceptance: July 14, 2025
- Camera ready: July 30, 2025
All submission deadlines are end-of-day in the Anywhere on Earth (AoE) time zone.