Generative Monte Carlo

Jian Huang

University of Iowa


Date: Wednesday, September 15, 2021
Time: 4:00 pm - 5:00 pm

Abstract: Learning a probability distribution based on a random sample and sampling from a given distribution are two fundamental problems in statistics and machine learning. These two problems have been studied intensively as two separate questions in the literature. In recent years, generative learning approaches such as generative adversarial networks have been proven to be effective in learning distributions of high-dimensional complex data. In this talk, we consider the problems of sampling from conditional and unnormalized distribution and illustrate that learning and sampling are two sides of the same coin in the context of generative learning. The key to the success of such an approach is the power of deep neural networks in approximating high-dimensional functions. Numerical experiments on benchmark image data and simulation from multimodal distributions demonstrate that generative Monte Carlo has the potential to be a useful addition to the methods for learning and sampling.


*All talks will be held virtually through Zoom.  Please contact for talk links.*