Welcome to the Uncertainty Quantification & Scientific Machine Learning Group!
Our research focuses on developing computational methods of uncertainty quantification and machine learning for complex systems in science, engineering, and medicine. We seek to derive these methods under a rigorous framework of mathematics and statistics together with computational feasibility and scalability. Our projects entail answering questions such as:
- How much uncertainty accompanies the model prediction, and how can we reduce it from new data/evidence?
- What data should we acquire next, and how many? (See our review paper on optimal experimental design!)
- How can machine learning be used together with physical modeling?
News
- (2024/09) MORe
- Thomas Coons is giving a presentation “Adaptive Covariance Estimation via Surrogate Modeling for Multi-fidelity Estimation”.
- Aniket Jivani is presenting a poster “Propagation of Uncertainties in Data-driven Learning of ODEs”.
- (2024/09) Hongfan Chen is presenting a poster “Hypothesis testing of gait measure differences between at-home and in-clinic settings” at the e-HAIL Symposium.
- (2024/09) Yossi Cohen is presenting a poster “Inverted Transformers for Effective Post-Calibration in
Multi-Site Wind Power Forecasts” at the MIDAS Postdoctoral Fellows AI Bootcamp. - (2024/09) Xun Huan is giving a presentation “Bayesian Optimal Experimental Design” at the University of Michigan Applied Physics Seminar.
- (2024/09) Our paper “Optimal experimental design: Formulations and computations” is published in Acta Numerica. See arXiv for a more up-to-date version!
- (2024/08) Our preprint “A likelihood-free approach to goal-oriented Bayesian optimal experimental design” is available on arXiv.
- (2024/07) Our paper “Shapley-based explainable AI for clustering applications in fault diagnosis and prognosis” is published in Journal of Intelligent Manufacturing.
- (2024/07) Our paper “Uncertainty quantification of graph convolution neural network models of evolving processes” is published in Computer Methods in Applied Mechanics and Engineering.
- (2024/06) Our paper “Probabilistic machine learning for battery health diagnostics and prognostics—Review and perspectives” is published in npj Materials Sustainability.
- (2024/04) Our preprint “Variational Bayesian optimal experimental design with normalizing flows” is available on arXiv.
- Upcoming/Interesting Events:
- (2024/10) Yossi Cohen is giving a presentation “Explainable Machine Learning for Knowledge Discovery in Advanced Manufacturing Systems” at INFORMS.
- (2024/10) Xun Huan is giving a presentation “Fokker–Planck-Based Inverse Reinforcement Learning” at SIAM MDS.
- (2024/12) AGU
- (2025/03) SIAM CSE
- (2025/10) IMSI Workshop on Optimal Control and Decision Making Under Uncertainty for Digital Twins
We are grateful to all our current and past sponsors.
