Camera-ready deadline: Friday,June 26th (AOE).
Call for Papers
Structured Data for Health (SD4H) workshop at ICML 2026 invites submissions focused on the unique challenges and opportunities in modeling structured data to advance human health. Structured data is ubiquitous in modern healthcare, spanning tabular electronic health records, high-frequency physiological time series, and irregular clinical measurements collected over time. This workshop brings together researchers working across machine learning, clinical informatics, and wearable sensing to address questions at the intersection of structured data modeling and healthcare, with the goal of enabling holistic, interpretable, and clinically impactful methods.
We invite you to submit your work by May 2, 2026 (Anywhere on Earth).
Topics of Interest
Submissions may explore but are not limited to:
Models & Methods
- Novel Architectures: State-space models, diffusion models, etc.
- Foundation Models: Pre-training, scaling, and alignment.
- Deep Learning: Supervised, self-supervised, and unsupervised methods.
- Probabilistic Modeling: Uncertainty quantification and Bayesian methods.
- Sequential Decision-Making: Reinforcement learning and optimal control.
Data-Specific Challenges
- Multimodal Learning: Fusing structured data (e.g. time-series, tables, graphics) with images or text.
- Structured Adaptation: Using LLMs or other methods to introduce structure to unorganized data (i.e. free-text).
- Irregular & Missing Data: Handling sparse or irregularly-sampled series.
- Complex Signals: Modeling high-dimensional or multi-resolution data.
- Causal Inference: Inferring cause-and-effect from observational data.
- Representation Learning & Adaptation: Pre-training, adaptation strategies that transfer across patients, devices, test-time adaptation.
Applications & Trustworthy AI
- Clinical Applications: Forecasting, risk stratification, digital biomarkers.
- Trust & Reliability: Explainability, fairness, robustness, and privacy.
- Deployment & Implementation: Real-world case studies and MLOps, federated evaluation, online learning during deployment.
- New Resources: Public datasets, benchmarks, and software.
Submission Instructions
We invite short papers of up to 4 pages, excluding references and appendices. All submissions should follow the official ICML'26 paper formatting guidelines (paper checklist is NOT needed) and be fully anonymized for a double-blind review process. Submitted work should be original and unpublished, though submissions on preprint servers like arXiv are permitted. Authors will be asked to confirm that their submissions accord with the ICML Code of Conduct. Note that papers not conforming to these instructions will be desk rejected.
All accepted papers must be presented in person at the workshop. This is a non-archival venue and there will be no formal proceedings.
Important Dates
Submission Deadline: 2 May 2026
Workshop Date: July 10/11, 2026 (Co-located with ICML'26 in Seoul, South Korea)
Frequently Asked Questions (FAQ)
- Q1. What does “non-archival venue” mean?
- A non-archival venue is one where accepted papers are not formally published in proceedings. This means the work will not appear in a permanent, citable archive.
- Q2. Does this mean accepted papers can be submitted elsewhere (e.g., to a journal or another conference)?
- Yes! Since our workshop is non-archival, you are free to submit your work elsewhere. We do recommend checking the policies of your target journal or conference, as some may have restrictions on prior presentations — even at non-archival events.
- Q3. Can I submit a paper that has been withdrawn from the main track of a conference?
- Absolutely. You are welcome to submit such work to our workshop.
- Q4. Can I submit the same work to your workshop and another venue (e.g., other ICML workshop) at the same time?
- Yes, concurrent submissions are allowed. Please ensure that your other target venue’s rules permit this.
- Q5. Can I present work at the workshop that has already been accepted elsewhere?
- Yes, provided that the other venue’s policies allow for prior or concurrent presentation at a non-archival workshop. Always confirm with the other venue first.
Sponsors
Bronze Level
