Uncertainty Quantification Lab’s mission is to accelerate discoveries and decision-making under uncertainty through novel computational tools based on a deep integration of computation, modeling, and experimentation.

Machine Learning in Catalysis

Machine learning models can help to reduce the large computational cost involved in computing various adsorption and transition-state energies of all possible surface states on a large number of catalyst models, and overcome the shortcomings of linear scaling relations for more complex chemistries.

Aflatoxin Prediction in Corn

The goal of this project is to develop a general predictive modeling framework for calculating mycotoxin incidence in US crop fields. Prediction and control of the most potent carcinogenic mycotoxin, aflatoxin, is a fundamental challenge for US grain industry, poultry producers, and makers of dairy products.

Uncertainty Quantification in Deep Learning

Exact Bayesian neural network methods are intractable and non-applicable for real-world applications. We propose an approximate estimation of the weights uncertainty using Ensemble Kalman Filter, which is easily scalable to a large number of weights.

Uncertainty Quantification (UQ) is the theoretical and computational fabric that connects the three pillars of science – theory, experimentation, and computation – through which uncertainties are characterized and informed to guide the scientific discovery and decision-making process.

The central challenge in using computational models for scientific discovery, engineering design, or decision support is that the process follows a path contaminated with errors and uncertainties. The inherent uncertainties are the result of many factors: experimental uncertainties, model structure inadequacies, uncertainties in model parameters and initial conditions, as well as errors due to sampling and numerical discretization.

The UQ research activities in our group fall into two categories: development of methodologies and algorithms and applications via funded collaborations with researchers in various sciences and engineering disciplines. Every project is a mixture between methodology development and application, and as such, every graduate student in the UQ lab is exposed to both methodology and application.