Semi-supervised Elastic Manifold Embedding with Deep Learning Architecture

Abstract Graph-based embedding aims to reduce the dimension of high dimensional data and to extract relevant features for learning tasks. In this letter, we propose an Elastic graph-based embedding with deep architecture which deeply explores the structural information of the data. We introduce a flexible deep learning that can overcome the limitations and weaknesses of single-layer learning models. The proposed deep architecture incorporates the geometrical manifold structure of the data. The resulting framework can be used for semi-supervised and supervised settings. Besides, the resulting optimization problems can be solved efficiently. We apply the algorithm on five public image datasets including scene, face and object datasets. These experiments demonstrate the effectiveness of the proposed embedding method, and also show that the proposed method compares favorably with many competing state-of-the-art graph-based methods.