What Matters in Your Model? Uncertainty Quantification and Sensitivity Analysis

Mathematical models are simplified representations of complex systems, relying on uncertain inputs, boundary conditions, and model parameters. Uncertainty is therefore inherent to modelling, yet predictions are increasingly used to inform decisions in contexts where it cannot simply be eliminated. The key challenge is not to remove uncertainty, but to understand, quantify, and manage it well enough to use model predictions responsibly.

Uncertainty quantification (UQ) asks: How uncertain are our predictions? Sensitivity analysis (SA) asks: Which factors drive that uncertainty, and which matter most? Together, they provide a framework for understanding model behaviour, identifying influential parameters, prioritising data collection, and directing modelling and validation efforts where they can have the greatest impact.

In this webinar, Prof. Leif Rune Hellevik from the Norwegian University of Science and Technology (NTNU) will introduce the motivation and key concepts behind UQ and global sensitivity analysis, using examples from cardiovascular modelling. The session will explore practical approaches, including Monte Carlo methods, polynomial chaos expansions, and interactive Jupyter notebooks, and discuss how UQ and SA can support credible model-based predictions and more informed decision-making.

The webinar will be moderated by the member of the VPH Young Scientists Committee, Dr. Rahul Manoj.

REGISTER HERE

Speaker’s bio:

Leif Rune Hellevik

Leif Rune Hellevik is a Professor in the Division of Biomechanics, Department of Structural Engineering, at the Norwegian University of Science and Technology (NTNU), Trondheim, Norway. He served as Vice Dean for master’s programmes at the Faculty of Engineering from 2017 to 2024. His research spans cardiovascular biomechanics, wave propagation, fluid-structure interaction, mathematical modelling, and scientific computing. Since 2011, he has led the development of STARFiSh (Stochastic Arterial Flow Simulations), a framework for modelling arterial pressure and flow waves with uncertainty quantification and sensitivity analysis. His group has also developed physics-based and data-driven reduced-order models for estimating fractional flow reserve from coronary CT angiography (CT-FFR). More recently, he has contributed to My Medical Digital Twin (MyMDT), which integrates exercise physiology, hypertension research, mathematical modelling, machine learning, biosensors, and genetic epidemiology to develop personalised digital twins for hypertension prevention and treatment. He also led the NTNU initiative IKTiSU (2013–2015) and served as an Executive Director of the VPH Society from 2024 to 2026.

Rahul Manoj

Rahul Manoj is a Senior R&D Engineer in the Technology Research Group at Orvue Surgical A/S, Stavanger, Norway. His research focuses on translational biomedical engineering, integrating physiology, clinical measurements, sensing technologies, and interpretable computational models to support clinical decision-making. From 2024 to 2026, he was a Postdoctoral Research Fellow in the Biomechanics Group at the Norwegian University of Science and Technology (NTNU), Trondheim, where he combined echocardiography, shear wave elastography, and hybrid models to assess diastolic dysfunction and their verification and validation. He received his M.S. and Ph.D. in Electrical Engineering from the Indian Institute of Technology (IIT) Madras, India, in 2024, and his B.Tech. in Electrical and Electronics Engineering from the National Institute of Technology (NIT) Calicut, India, in 2016. His doctoral research focused on early vascular ageing and central blood pressure measurement.