Ee13-Data Science and Artificial Intelligence for Medicine and Health (CDIAMS)
Epidemiology and Public Health / IBS-Ee13 / Emerging
The general objective of the group focuses on the application and development of XAI techniques and IoT infrastructures to generate accurate and understandable descriptive or predictive models from heterogeneous and complex data related to three lines of health research, allowing health professionals to carry out more accurate and detailed analyses of their patients in order to provide them with more effective and personalized treatments that help prevent diseases, having a significant impact on people's health and well-being. This general objective can be divided into the following sub-objectives:
A. Childhood obesity:
- Develop and apply advanced supervised ML techniques that generate understandable models from omic, clinical and exposome data from different life stages through efficient learning and fusion of information sources and addressing current challenges, allowing the design of decision support systems (DSS) that improve the ability to predict the development of IR and/or obesity or a clinical parameter of interest at the primary care level.
- Develop and apply advanced unsupervised ML techniques to identify subgroups and extract association patterns and rules, integrating multiomic, clinical and environmental data, to identify molecular inter-omic networks or epistasis phenomena underlying obesity and IR in the pediatric population, and that allow us to detect interesting associations with clinical and exposomic data.
B. Prostate cancer: Develop advanced supervised ML techniques to generate understandable models and make use of post-hoc explainability techniques that provide explanations about the models generated from different sources of information (gene expression levels, clinical levels, etc.) to identify and evaluate gene signatures that allow clinicians to classify and/or determine the aggressiveness of the tumor in combination with the data obtained from the microscope.
C. Indoor air management for respiratory diseases: Develop an open-access software and hardware infrastructure to provide an IoT solution to manage air quality in the home, using controllers based on understandable models that allow the management of heterogeneous and printed data and follow various standards (such as IEC 61131 and IEEE Std 1855-2016) to facilitate migration across multiple hardware platforms and the reuse of developed applications.
Research lines
- AI for early diagnosis, disease progression, risk assessment and personalized treatment.
- AI for real-time monitoring of patients with complex diseases.