We formulated the existing two-stage SSL problem into a unified optimization framework – termed as Self-Taught Semi-Supervised Learning (STSSL), in which both the given labels and the estimated labels are incorporated to refine the affinity matrix and to facilitate the unknown label estimation.
We propose a novel regularized Semi-Supervised Latent Dirichlet Allocation (r-SSLDA) for learning visual concept classifiers.
This paper proposed a novel framework, named fixedrank representation (FRR), for robust unsupervised visual learning.
This paper proposes a novel informative graph, called the nonnegative low rank and sparse graph (NNLRS-graph), for graph-based semi-supervised learning.