The document summarizes a survey paper on private machine learning. It discusses how privacy has become an emerging issue with increased data collection and more sophisticated attackers. The paper surveys current private machine learning attacks targeting training data, models, and inferences. It also reviews defense mechanisms like encryption, obfuscation, and aggregation. The future outlook is that attackers will continue developing new techniques while researchers work to protect privacy through methods such as noise addition, obfuscation, encryption, and generative models.