![Data-Driven Decision Processes_hi-res logo](/sites/default/files/styles/workshop_banner_sm_1x/public/2023-03/Data-Driven%20Decision%20Processes_hi-res.png.jpg?itok=WR1qG1Qx)
Abstract
In this talk we will discuss online decision-making problems where the input is partly adversarial and partly stochastic. In keeping with the theme of competitive analysis, we will compare the performance of the online algorithm against a hindsight optimum that observes the entire input before making decisions. I will survey techniques and results for the two dominant paradigms for these settings, namely secretary problems and prophet inequalities.