Transfer Learning with Graph Co-Regularization

Transfer learning proves to be effective for leveraging labeled data in the source domain build an accurate classifier target domain. The basic assumption behind transfer is that involved domains share some common latent factors. Previous methods usually explore these factors by optimizing two separate objective functions, i.e., either maximizing empirical likelihood, or preserving geometric structure. Actually, functions are complementary each other and them simultaneously can make solution smoother further improve accuracy of final model. In this paper, we propose a novel approach called Graph co-regularized Learning (GTL) purpose, which integrates seamlessly into one unified optimization problem. Thereafter, present iterative algorithm problem with rigorous analysis on convergence complexity. Our study open sets validates GTL consistently classification compared state-of-the-art methods.