About the Group
AI 4 Continuous Learning is dedicated to developing and applying adaptive artificial intelligence systems to study ageing as a continuous, lifelong process. The group investigates how continuous, lifelong learning models can capture dynamic, nonlinear changes in biological, clinical, and functional states across the lifespan.
By treating ageing as a time-evolving phenomenon, the group moves beyond static disease models, enabling AI systems that learn from longitudinal and multimodal data while preserving knowledge over time.
Vision & Impact
AI 4 Continuous Learning contributes to MIA-Portugal's mission by positioning AI as a core enabler for understanding ageing across the lifespan. By developing adaptive, interpretable, and privacy-preserving AI systems, the group aims to support early prevention, personalised interventions, and healthier ageing at both individual and population levels
Mission and Research Aims
The mission of AI 4 Continuous Learning is to advance AI-driven ageing research by designing lifelong, interpretable, and trustworthy learning systems. The group aims to:
- Model ageing as a continuous adaptive process rather than a set of isolated conditions.
- Develop a continuous, lifelong-learning AI capable of evolving with new data, settings, and populations.
- Integrate AI with biomedical, clinical, and population-level ageing data.
- Support personalised prevention and intervention strategies for healthy ageing.
- Promote responsible, transparent, and human-centric AI in biomedical research.
Research Areas
- The group conducts interdisciplinary research at the intersection of artificial intelligence, ageing research, and precision medicine. Core research areas include:
- Continuous & Lifelong Learning AI
- Development of adaptive models that learn over time without catastrophic forgetting, enabling robust longitudinal modelling.
- AI-Based Modelling of Ageing Trajectories
- Representation of ageing as a dynamic process across molecular, physiological, behavioural, and clinical scales
- Distributed & Privacy-Preserving Learning
- Distributed learning frameworks for collaborative research across institutions while preserving data privacy.
- Interpretable & Trusted AI
- Explainable and reliable models to support biological insight, clinical understanding, and decision-making.
- AI for Prevention & Intervention Design
- Data-driven simulation and optimisation of personalised interventions and timing to improve long-term outcomes.
Selected Publications
apala, K., Tworek, P. & Sousa, J. (2025). Stability of Machine Learning Predictive Features Under Limited Data. IEEE Transactions on Knowledge and Data Engineering, PP(99), 1–13. https://doi.org/10.1109/TKDE.2025.3580671
* Gherardini, L., Tworek, P., Szczypka, M., Khan, Y., Mikołajczyk, M., Lewandowski, R. & Sousa, J. (2025). Artificial Intelligence in Medicine. 23rd International Conference, AIME 2025, Proceedings, Part II, 171–175. https://doi.org/10.1007/978-3-031-95841-0_32
* Tworek, P., Szczypka, M., Kahan, J., Mikołajczyk, M., Lewandowski, R. & Sousa, J. (2025). Artificial Intelligence in Medicine. 23rd International Conference, AIME 2025, Proceedings, Part I, 448–456. https://doi.org/10.1007/978-3-031-95838-0_44
* Gherardini, L., Varma, V. R., Capała, K., Woods, R. & Sousa, J. (2024). CACTUS: A Comprehensive Abstraction and Classification Tool for Uncovering Structures. ACM Transactions on Intelligent Systems and Technology, 15(3), 1–23. https://doi.org/10.1145/3649459
* Lyall, D. M., Kormilitzin, A., Lancaster, C., Sousa, J., Petermann-Rocha, F., Buckley, C., Harshfield, E. L., Iveson, M. H., Madan, C. R., McArdle, R., Newby, D., Orgeta, V., Tang, E., Tamburin, S., Thakur, L. S., Lourida, I., Network, T. D. D. P. (DEMON), Llewellyn, D. J. & Ranson, J. M. (2023). Artificial intelligence for dementia—Applied models and digital health. Alzheimer's & Dementia. https://doi.org/10.1002/alz.13391
* Ibias, A., Varma, V. R., Capała, K., Gherardini, L. & Sousa, J. (2023). SaNDA: A Small and iNcomplete Dataset Analyser. Information Sciences, 640, 119078. https://doi.org/10.1016/j.ins.2023.119078