Knowledge Transfer in Principal Component Analysis Studies
The author proposes a two-step transfer learning algorithm for unsupervised tasks in PCA studies, emphasizing the importance of shared subspace information and the gain from knowledge transfer.