Representation Learning for Medical Time-Series
Representation Learning for Medical Time-Series
We develop models that learn robust, informative representations of longitudinal health data. Using graph neural networks, transformers, few-shot, and multimodal learning, we capture temporal and relational patterns in patient records, wearable signals, and clinical notes.
Publications
- Zhu, Mingcheng, et al. "Bridging data gaps of rare conditions in ICU: a multi-disease adaptation approach for clinical prediction." npj Digital Medicine (2026). paper, code
- Cai, Zi, et al. "ProtoEHR: Hierarchical Prototype Learning for EHR-based Healthcare Predictions." Proceedings of the 34th ACM International Conference on Information and Knowledge Management. 2025. paper, code
- Li, Chenqi, et al. "AnchorInv: Few-Shot Class-Incremental Learning of Physiological Signals via Feature Space-Guided Inversion." Proceedings of the AAAI Conference on Artificial Intelligence. Vol. 39. No. 13. 2025. paper, code
- Zhu, Mingcheng, et al. "MedTPE: Compressing Long EHR Sequence for LLM-based Clinical Prediction with Token-Pair Encoding." KDD 2025 MILETS Workshop, 2025. paper