Dr. Mu Zhu, Professor of Statistics at the University of Waterloo, Canada will discuss his work with generative nural networks.
From Dr. Zhu: "I will share some examples of what we have recently been able to do for statistics with generative neural networks. We are able to generate quasi Monte Carlo samples from practically "any" multivariate distribution as long as we have some training data from it. We can forecast multivariate time series without restricting ourselves to only a few parametric copula families for describing the underlying cross-sectional dependence. We have also created a visually-based diagnostic tool for model selection. All three projects are joint work with Marius Hofert and Avinash Prasad."
Dr. Mu Zhu is a Professor of Statistics at the University of Waterloo, Canada, and a Fellow of the American Statistical Association. He is the Director of the Data Science Program in the Department of Statistics and Actuarial Science, University of Waterloo, Canada. He obtained his undergraduate degree from Harvard University in 1995, and his PhD in Statistics from Stanford University in 2001. Professor Zhu’s research works include efficient kernel machines for rare target detection and ensemble methods for variable selection, algorithms for making personalized recommendations, dependent modeling, generative neural networks and applications of machine learning , including healthcare informatics, e-commerce, sports analytics, dependent modeling in actuarial analytics, and others.
Pearce Hall 138
|Sponsor:||College of Science & Engineering|
|Contact:||Shari Jackson email@example.com 989-774-7464|
|Register for this event|