MSc in Applied Statistical Modelling & Health Informatics at King's College London
Research interests: Clinical NLP, multimodal EHR learning, trustworthy health AI, and benchmark design.
TIMELY-Bench
A time-aware clinical benchmark for multimodal ICU prediction using MIMIC-IV.
This project studies temporal alignment, structured–text fusion, and data leakage in clinical prediction pipelines, with experiments conducted on KCL CREATE HPC.
Submitted to HealTAC 2026
→ Code repository will be released after publication.
HalluCXR
A benchmark for evaluating hallucination in vision-language models for chest radiograph interpretation.
This project evaluates six VLMs and explores ensemble-based mitigation strategies with human annotation.
Submitted to CVPR 2026 AI4RWC Workshop
→ github.com/haoyu-haoyu/HalluCXR
Technical Areas
Clinical NLP Multimodal Learning Time-series Modelling Trustworthy AI EHR Representation Learning
Tools & Frameworks
PyTorch Transformers XGBoost scikit-learn spaCy
Programming
Python R SQL TypeScript Go Shell
Clinical Data & Infrastructure
MIMIC-III/IV OMOP CDM FHIR DICOM Docker SLURM/HPC
Actively seeking PhD opportunities (2026–2027) in Clinical NLP, Health AI, and Multimodal EHR Research.
📧 haoyu.7.wang@kcl.ac.uk ·
🌐 GitHub