Bayesian Semi-supervised Multicategory Classification under Nonparanormality

Attend this 57-minute seminar exploring Bayesian approaches to semi-supervised multicategory classification problems under nonparanormal conditions. Learn from Professor Subhashis Ghoshal of North Carolina State University as he presents advanced statistical methods for handling classification tasks when dealing with partially labeled data and non-normal distributions. Discover how Bayesian frameworks can be applied to multicategory classification scenarios where traditional parametric assumptions may not hold, particularly focusing on nonparanormal distributions that allow for more flexible modeling of real-world data. Gain insights into semi-supervised learning techniques that leverage both labeled and unlabeled data to improve classification performance, while understanding the theoretical foundations and practical implications of these methods. Explore the intersection of Bayesian statistics and machine learning in the context of uncertainty quantification and prediction calibration, as part of the broader research programme on representing, calibrating, and leveraging prediction uncertainty from statistics to machine learning.