LLM Agent for Capacity Modelling in Health Care Provision

This PhD project is based in the AI Centre for Doctoral Training in Biomedical Innovation (AI4BI CDT), a part of the University of Edinburgh, funded by the UK Research and Innovation. It provides fully funded PhD studentships to UK, EU, and international candidates. Link: https://www.ai4biomed.io/.

This project is in collaboration with Brigham and Women’s Hospital, a top hospital in the United States, and a premier teaching hospital of Harvard Medical School. Link: https://www.brighamandwomens.org/.

This project aims to design AI agents based on large language models (LLMs) to tackle the capacity modelling challenges in health care provision: (1) resilience: ensuring resilience in the face of public health crises, such as unexpected surges in demand, or unforeseen reduction in supplies, (2) efficiency: enhancing patient outcomes, minimising delays in care, reducing waiting times, and ensuring timely access to medical services, and (3) ethics: ensuring equitable resource allocation, transparency in decision-making, patient autonomy, etc.

The research will develop a comprehensive suite of models and optimisation algorithms for LLM agents specifically designed to address capacity modelling in health care provision, offering mathematical guarantees. These models will be integrated into a prototype system capable of functioning in real-world health care scenarios, such as hospital resource allocation, patient scheduling, and service demand prediction, ensuring practical applicability.

It will be an interdisciplinary research project – related techniques including LLM, multi-objective constrained optimisation, reinforcement learning, algorithmic fairness, mechanism design, algorithmic game theory, learning theory, optimisation theory, etc., whilst domain knowledge in health care system, especially the capacity modelling of health care provision, is crucial.

Candidate’s profile

• A good Bachelor’s degree (First Class Honours or international equivalent) and/or Master’s degree in a relevant subject (mathematics, statistics, economics, or related subject).
• Strong programming skills. Experiences of software development and programming competition are highly desirable.
• A strong mathematical background, with an emphasis on analysis, probability, and statistics. Recipients of mathematics competition medals are desirable.
• Proficiency in English (both oral and written).
• Relevant research experiences in machine learning, statistics, etc. are desirable.

Lead Supervisor:

Fengxiang He, School of Informatics, University of Edinburgh, F.He@ed.ac.uk, https://fengxianghe.github.io/

Co-supervisors:

Filippo Menolascina, School of Engineering, University of Edinburgh, Filippo.Menolascina@ed.ac.uk

Lisa Lehmann, Brigham and Women’s Hospital & Harvard Medical School, lzslehmann@gmail.com

Contact

Applicants are highly encouraged to contact Fengxiang He (F.He@ed.ac.uk) to discuss your case.

Environment

The University of Edinburgh is constantly ranked among the world’s top universities and is a highly international environment with several centres of excellence. The School of Informatics is one of the largest in Europe and currently the top Informatics institute in the UK for research power, with 40% of its research outputs considered world-leading (top grade), and almost 50% considered top grade for societal impact. The School of Informatics is exceptionally strong in the area of AI and Theoretical Computer Science, hosting one of the largest group for AI and Foundations of Computer Science in the world. The successful applicant will be part of the Artificial Intelligence and its Applications Institute and will have the opportunity to interact with the other members of the group and more widely within the School of Informatics.

Project webpage: ‘Developing LLM Agents for Resilient, Efficient, and Ethical Capacity Modelling in Health Care Provision’, https://www.ai4biomed.io/research/projects-2025/

Application information: https://www.ai4biomed.io/how-to-apply/