Abstract:

Semi-supervised learning attempts to take advantage of the large amount of unlabeled information present in many datasets. However, unlabeled data will often contain samples outside the classes of interest. Many existing semi-supervised learning methods do not address this issue. In this work, we propose to address this problem. Generative Adversarial Networks (GANs) are a state of the art deep learning architecture that has been used in the literature for semi-supervised classification. We propose to modify this architecture by adding a Self-Organizing Map to allow it to handle outliers in the unlabeled dataset. This document will describe the modified architecture, as well as provide results of its performance on several different datasets.

Links:

Citation:

R. Fick, "Robust GANs for Semi-Supervised Classification," Ph.D Thesis, Gainesville, FL, 2022.

@phdthesis{Fick2022Thesis,
author = {Ronald Fick},
title = {Robust GANs for Semi-Supervised Classification},
school = {Univ. of Florida},
year = {2022},
address = {Gainesville, FL},
}