Uncertainty Quantification Group

Recent publications

For the complete list of articles see our Google scholar page

Predictive analytics for smart cities

Online predictive model for post-disaster traffic conditions

An online algorithm for traffic flow prediction that requires minimal computational capacity, successfully tested on the Woolsey fire in California, following which traffic patterns significantly changed.

Learn more

Recursive Traffic Flow Predictions with Faulty Sensors

A predictive model for traffic flows at locations with faulty sensors, solely by using traffic measurements from neighboring sensors.

Learn more

Machine learning-based optimization and control

Fast reliability analysis for bridge networks

A deep learning framework to faciliate seismic analysis of bridge networks.

Learn more

Surrogate-based probabilistic control of power systems

A fast surrogate-based probabilistic voltage control for power systems with many distributed generation units.

Learn more

Physics-based deep learning

Deep learning solution for random PDEs

A new solution approach for high dimensional random PDEs based on a feed-forward fully-connected deep residual network, with either strong or weak enforcement of initial and boundary constraints.

Learn more

Physics-driven regularization of deep neural networks

A DNN training approach that utilizes the known governing laws (in the form of PDEs) and regularizes DNN models by penalizing divergence form those laws.

Learn more

Connected and autonomous vehicles

Efficient data collection from connected vehicles

A collection scheme that selects and transmits only a small subset of data to alleviate data transmission burden and still meet precision requirement, successfully tested on the 10,000 connected vehicle trips in the Safety Pilot Model Deployment dataset.

Learn more