Corelab Seminar

Alkis Kalavasis
On Learning and Covering Sums of Independent Integer Random Variables

We will talk about the problem of learning sums of independent and integer random variables, which falls in the agenda of distribution learning. We are going to start with an overview of the area, introduce the main learning task and discuss the previous work on this important learning theory problem. Next, we will talk about our contribution on learning and covering new families of distributions, bypassing the existing learning lower bounds established by De, Long and Servedio at FOCS '18.