This document summarizes a presentation on Locality Preserving Projections (LPP), a dimensionality reduction technique. LPP preserves the local neighborhood structure of the data when projecting it to a lower-dimensional space. It constructs an adjacency graph representing the connections between nearby data points, chooses weights for these connections, and computes eigenmaps to perform the dimensionality reduction while maintaining locality. LPP is compared to Principal Component Analysis (PCA), the most common dimensionality reduction method, which aims to preserve global variance rather than local neighborhood structure.