Intelligent Decision Support

Effective decision support is not only about predicting the best action. Particularly in healthcare and long-term care, recommendations must reconcile professional knowledge and clinical effectiveness with each person’s lifestyle, preferences, and everyday constraints.

We use ensemble learning, confidence-aware adaptive inference, and hierarchical multi-agent decision-making to integrate heterogeneous evidence, manage uncertainty, and determine how limited resources should be used. These methods support reliable decisions when measurements are noisy, computation is constrained, or multiple objectives interact.

Our work includes spatio-temporal gait estimation, power-efficient sleep-stage classification on wearable devices, and uncertainty-aware planning. Building on these foundations, we aim to develop personalized care support that provides recommendations that are both professionally sound and practical in daily life.