Area of Expertise

Main knowledge contributions towards intelligent data analytics fall into a range of areas including:

-   Novel algorithms for causal discovery that integrate structural causal models (SCMs) with generative processes to better capture how real-world data are produced. They are grounded in the principle that SCMs provide the fundamental framework for understanding data generation, ensuring interpretability and theoretical soundness. By combining causal structure discovery with the expressive power of generative AI, the work bridges the gap between theory and practical data-driven modelling to open up a new pathway for discovering reliable causal knowledge directly from complex, high-dimensional real-world datasets. We further apply this framework to healthcare domains, including emulating clinical trials to extrapolate their results for broader patient populations.

- Knowledge discovery in AI for healthcare to support patient self-management of general health and chronic conditions, involving smart monitoring, data validation from heterogeneous sensors,  personal activity and event recognition,  health information recommendation, personal health status estimation and serious gaming.

-    Intelligent data analytics for computational creativity in AI by coordinating the EC-funded Dr Inventor research project and leading the development of the Dr Inventor platform. The Dr Inventor surrogate acts as a personal research assistant, utilising machine-empowered search and computation to bring researchers extended perspectives for scientific innovation by informing them of a broad spectrum of relevant research concepts and approaches, by assessing the novelty of research ideas, and by offering suggestions of new concepts and workflows with unexpected features for new scientific discovery.

-   Visualization and parallel computing (GPU) for large-scale medical data, , including transfer function for feature enhancement in volume rendering of medical data; viewpoint selection and lighting design for volume rendering of medical data; Non-photorealistic volume rendering for feature enhancement from medical data; GPU-based iso-surface extraction from volume data and automated GPU-based parallelisation for images operations and image feature extractions.

-   Visual analytics for health data to support the navigation, query and understanding of health records, clinical driven research in predictive models for cancer growth in response to treatment options, and the discovery of data patterns within patient cohort in both clinical and lifestyle domains

-   Computer vision and machine learning for computer graphics research, including sparse modelling and representation for human motions,  blind motion deblur for natural images,  adaptive texture synthesis for high fidelity images and image based rendering based on inferences in machine learning

-    Health data interoperability to support long-term collection of personal health information by aggregating  electronic and personal health records, lifestyle data and drug information in a decentralised approach to offer easy access to personal medical history, empower the patients, improve self-management, and facilitate clinical research with significant advantages in privacy, security, safety, transparency and data integrity.

The recent active research projects include: 

VisC: Causal Counterfactual visualisation for human causal decision making – A case study in healthcare. Principal Investigator, EPSRC, EP/X029778/1, July 2023 – Dec 2025

MRC-GAN: Virtual Clinical Trial Emulation with Generative AI Models, Principal Investigator,  MRC, MR/X005925/1, Sept 2022 – Feb 2023 

Qualifications

• PhD in Computer Science, Zhejiang University, China
• PGCERT Higher Education 
• The Higher Education Academy Fellow -