Two scientific challenges arise in numerous engineering fields: uncertainties and high-volume data. Uncertainties are unavoidable in almost all realistic engineering design, causing performance and yield loss in electronic and photonic integrated circuit dessign, collisions in autonomous systems, and system failures in energy systems. High-volume data appears in financial data analysis, medical imaging, scientific computing, AI-related tasks (e.g., deep learning) and many other domains. Processing high-volume data consumes massive computational and hardware resources.

​

In order to address the above challenges, we are investigating two correlated research topics: uncertainty quantification and tensor computation. We are interested in the following applications: (1) variation-aware electronic and photonic design automation, (2) hardware/algorithm co-optimization of machine learning and big data analytics, (3) modeling, control and verification of autonomous dynamic systems, (4) modeling and data analytics for MRI.​