Mingcheng Zhu

Me at Sagrada Familia

I am a full-time DPhil student at the AI4DH Lab, University of Oxford, supervised by Prof. Tingting Zhu. My research focuses on modelling electronic health records as event streams, with a particular emphasis on rare diseases, clinical prediction, and few-shot or zero-shot clinical AI.

Before Oxford, I worked as a Senior Research Assistant at the Digital Medicine Lab, Duke-NUS Medical School, supervised by Prof. Nan Liu, where I worked on clinical AI fairness and controllable representation learning. I received an MRes from Imperial College London in 2023 and a BEng from Sichuan University in 2022.

My current interests include efficient large language models for healthcare, EHR event stream modelling, multimodal representation learning, embodied intelligence, algorithmic fairness, interpretability, and uncertainty estimation for high-stakes decision-making.

News

Jun 07, 2026	June 2026: Our work, “Attention-Guided Fair AI Modeling for Skin Cancer Diagnosis”, was accepted to npj Digital Medicine.
May 24, 2026	May 2026: I started as a Research Intern at Spirit AI, working on VLM training for humanoid robot manipulation.
Apr 30, 2026	May 2026: Our work “From Token to Token Pair: Efficient Prompt Compression for Large Language Models in Clinical Prediction” was accepted to ICML 2026.

Selected Publications

ICML

From Token to Token Pair: Efficient Prompt Compression for Large Language Models in Clinical Prediction

Mingcheng Zhu,Zhiyao Luo,Yu Liu, andTingting Zhu

In International Conference on Machine Learning, 2026

Paper Code Bib

@inproceedings{zhu2026tokenpair,
  title = {From Token to Token Pair: Efficient Prompt Compression for Large Language Models in Clinical Prediction},
  author = {Zhu, Mingcheng and Luo, Zhiyao and Liu, Yu and Zhu, Tingting},
  booktitle = {International Conference on Machine Learning},
  year = {2026},
}

npj DM

Attention-Guided Fair AI Modeling for Skin Cancer Diagnosis

Mingcheng Zhu,Mingxuan Liu,Han Yuan,Yilin Ning,Zhiyao Luo,Tingting Zhu, andNan Liu

npj Digital Medicine, 2026

arXiv Bib

@article{zhu2026fairskin,
  title = {Attention-Guided Fair AI Modeling for Skin Cancer Diagnosis},
  author = {Zhu, Mingcheng and Liu, Mingxuan and Yuan, Han and Ning, Yilin and Luo, Zhiyao and Zhu, Tingting and Liu, Nan},
  journal = {npj Digital Medicine},
  year = {2026},
}

npj DM

Bridging Data Gaps of Rare Conditions in ICU: A Multi-Disease Adaptation Approach for Clinical Prediction

Mingcheng Zhu,Yu Liu,Zhiyao Luo, andTingting Zhu

npj Digital Medicine, 2026

Paper Code Bib

@article{zhu2026raredisease,
  title = {Bridging Data Gaps of Rare Conditions in ICU: A Multi-Disease Adaptation Approach for Clinical Prediction},
  author = {Zhu, Mingcheng and Liu, Yu and Luo, Zhiyao and Zhu, Tingting},
  journal = {npj Digital Medicine},
  volume = {9},
  pages = {7},
  year = {2026},
}

ICASSP

Cross-Representation Benchmarking in Time-Series Electronic Health Records for Clinical Outcome Prediction

Tianyi Chen,Mingcheng Zhu,Zhiyao Luo, andTingting Zhu

In IEEE International Conference on Acoustics, Speech, and Signal Processing, 2026

MSc student supervision

Paper Bib

@inproceedings{chen2026crossrepresentation,
  title = {Cross-Representation Benchmarking in Time-Series Electronic Health Records for Clinical Outcome Prediction},
  author = {Chen, Tianyi and Zhu, Mingcheng and Luo, Zhiyao and Zhu, Tingting},
  booktitle = {IEEE International Conference on Acoustics, Speech, and Signal Processing},
  pages = {7076--7080},
  year = {2026},
  note = {MSc student supervision},
}