About the Project
This PhD project will develop artificial intelligence methods for analysing complex medical imaging and biomedical data. The project will address a defined clinical or biomedical problem using appropriate AI techniques, which may include multimodal learning, spatiotemporal modelling, generative AI, foundation models and interpretable AI. Applications will involve real-world data such as retinal imaging, ultrasound and physiological signals.
Artificial intelligence is increasingly used to analyse medical images and other biomedical data, but many clinically important problems remain challenging because real-world data can be heterogeneous, multimodal, longitudinal and incomplete. Effective AI methods therefore need to extract useful information from complex data while producing reliable and clinically meaningful outputs.
This PhD project will develop and evaluate new AI methods to address a defined problem in medical imaging and biomedical data analysis. The specific research question will be refined at the start of the PhD in consultation with the supervisory team, based on the candidate's background, available data and the clinical problem being investigated. This project may use one or more of the following AI approaches:
1. Multimodal Foundation Models and Vision-Language Intelligence. It can be used to integrate medical images, text and other clinical information through shared representations and cross-modal alignment, supporting tasks such as multimodal understanding, medical question answering, report generation and clinical reasoning.
2. Spatiotemporal and Longitudinal Learning. It can be used to model medical data acquired over time, including imaging sequences and irregular longitudinal observations, enabling the analysis of temporal patterns, disease progression and future clinical states.
3. Generative AI and World Models. It can be used to learn latent representations of biomedical data, model possible state transitions, generate missing or future observations, and explore disease progression or potential responses to clinical intervention.
4. Human-centred and Interpretable AI. It can be incorporated to make AI outputs more transparent and clinically meaningful, including visual evidence localisation, uncertainty quantification, verifiable predictions and approaches that support effective human-AI interaction.
The project will use selected real-world biomedical datasets, with potential applications including retinal imaging, ultrasound imaging and video, and physiological signals such as EEG and ECG. These data will provide clinically relevant testbeds for developing and validating the proposed AI methods.
The student will receive training in machine learning, computer vision, medical image analysis, scientific computing, experimental design and research communication. The initial stage of the PhD will involve identifying and refining a specific research question, reviewing the relevant literature and establishing baseline methods. Subsequent work will focus on developing and validating novel methodology, publishing research findings and progressively developing an independent research programme leading to the doctoral thesis.
We invite applications from highly motivated and talented PhD candidates interested in developing state-of-the-art AI solutions for healthcare. Applicants must hold/achieve a minimum of a merit at master's degree level (or international equivalent) in a science, mathematics or engineering discipline. Applicants without a master's qualification may be considered on an exceptional basis, provided they hold a first-class undergraduate degree. The English language requirements must also be met by the start of the PhD.
Please contact He Zhao (he.zhao@liverpool.ac.uk) to discuss prior to submitting an online application. Please insert [PhD_CSC_application] in your email subject.
Prepare your application documents following the guidance on this website.
Funding Notes
The University of Liverpool and China Scholarship Council Awards