Abstract: In many prediction problems, it is beneficial to obtain confidence estimates for the classification output. We consider the problem of estimating confidence sets in multiclass classification ...

Code for our paper "Transformers as Neural Augmentors: Class Conditional Sentence Generation with Variational Bayes". Abstract: Data augmentation methods for Natural Language Processing tasks are ...

Abstract: We propose a generative mixture model classifier that allows for the class conditional densities to be represented by mixtures having certain subsets of their components shared or common ...