A Local Similarity-Preserving Framework for Nonlinear Dimensionality Reduction with Neural Networks | Read Paper on Bytez