Advances in Differential Privacy and Differentially Private Machine Learning

There has been an explosion of research on differential privacy (DP) and itsvarious applications in recent years, ranging from novel variants andaccounting techniques in differential privacy to the thriving field ofdifferentially private machine learning (DPML) to newer implementations inpractice, like those by various companies and organisations such as censusbureaus. Most recent surveys focus on the applications of differential privacyin particular contexts like data publishing, specific machine learning tasks,analysis of unstructured data, location privacy, etc. This work thus seeks tofill the gap for a survey that primarily discusses recent developments in thetheory of differential privacy along with newer DP variants, viz. Renyi DP andConcentrated DP, novel mechanisms and techniques, and the theoreticaldevelopments in differentially private machine learning in proper detail. Inaddition, this survey discusses its applications to privacy-preserving machinelearning in practice and a few practical implementations of DP.