International Data Science Challenge 2026

🎯 The Mission: Math Meets Hope

We believe math isn’t just numbers on a whiteboard—it’s the ultimate toolkit for finding hope hidden in the data. Your directive? Mathematics for Hope in Healthcare. You will tackle real-world clinical dilemmas using cutting-edge machine learning, deep learning, and advanced statistical modeling.

💻 The Challenge: Code to Save Lives

Step into the shoes of an AI clinician! Healthcare impact often depends on spotting risk patterns in physiological signals and medical imaging before a human eye can. But biomedical data is messy, noisy, and highly sensitive. Your challenge is to build models that aren’t just highly accurate, but transparent, reproducible, and actually useful in the real world.

The Dataset Directives: Teams must use at least one of the official datasets listed below, hosted directly on PhysioNet. External labeled datasets are strictly prohibited. Everyone fights in the same arena, regardless of the dataset you pick.

Pick Your Poison (The Datasets):

Choose your modality and lock in your modeling objective (e.g., classification, segmentation, signal decoding, or risk prediction):

1. bigP3BCI (P300-based Brain–Computer Interface) 🧠
   • The Data: EEG signals and event markers (EDF+ format).
   • The Mission: P300 detection, BCI decoding, or cross-session generalization.
2. Brugada-HUCA (12-Lead ECG for Brugada Syndrome) ❤️
   • The Data: 12-lead ECG recordings (WFDB format).
   • The Mission: Classify Brugada syndrome vs. Normal controls.
3. Myocardial Perfusion SPECT ☢️
   • The Data: Cardiac nuclear imaging / Rest SPECT (DICOM + NIfTI masks).
   • The Mission: Medical image analysis, segmentation of the left ventricular wall, or preprocessing/QA.
4. HYGD (Hillel Yaffe Glaucoma Dataset) 👁️
   • The Data: Retinal fundus images (JPG + Labels.csv).
   • The Mission: Glaucomatous optic neuropathy (GON) detection and quality-aware modeling.

Your Final Payload (Deliverables):

• 🛠️ The Method: A validated pipeline with strict train/val/test splits.
• 🔍 The Insights: Interpretable outputs (e.g., saliency/heatmaps for images, feature importance).
• 🗣️ The Impact: A real-world discussion on ethics, limitations, and how your model sparks “hope” in clinics.
• 📦 The Repo: Beautifully documented preprocessing for absolute reproducibility.

🗓️ Roadmap to the Grand Final

• 🏁 25 Feb: System Online & Registration Opens
• ⏳ 10 Mar: Registration Lock-in Deadline
• 🎉 14 Mar: Intl. Mathematics Day Celebration & Workshops
• 📤 25 Mar: Stage 1 Deadline (Code & Pitch Drop)
• 🔥 11 Apr: Stage 2 Grand Final & Winner Reveal