In classification, semi-supervised learning occurs when a large amount of unlabeled data is available with only a small number of labeled data. In this talk, I will discuss how to combine unlabeled and labeled data to enhance the generalization accuracy of classification. A large margin technique will be presented, which utilizes grouping information from unlabeled data, together with the concept of margins, in a form of regularization controlling the interplay between labeled and unlabeled data. Computational aspects will be discussed through difference convex programming, in addition to a tuning method that involves both labeled and unlabeled data, for tuning in regularization. Finally, numerical examples will be provided.

This work is joint with Junhui Wang.