Given two investors: one investing all of their annual target investment in January, and the other investing at some subannual frequency (e.g. monthly, quarterly, etc), which one wins at the end of the investments’ time-horizon? ...
Machine learning models depend on large troves of data to develop and improve their inference / prediction capabilities. These models generalize their abilities, but inevitably some of their training data is encoded or memorized within their trained parameters. This post explores a method of quantifying model memorization, factors that make models more prone to memorization, and provides some mitigations against memorization. ...
This post serves as a primer to Differential Privacy: presenting an intuitive foundation for the definition, proceeding to some math, and finally presenting a case study to demonstrate some key concepts. ...