This document summarizes a presentation on Factorization Machines and Neural Factorization Machines. It begins with an overview of Factorization Machines, describing them as a generic approach that can mimic many factorization models through feature engineering. It then discusses how FM combines the generality of feature engineering with the power of factorization models to model interactions between categorical variables, working well on sparse data. The document then introduces Neural Factorization Machines as an extension of FM to address its limitations, using a multi-layer feedforward neural network as the core component. It concludes by comparing FM and NFM and listing references.