Algorithmic stability for adaptive data analysis

Adaptivity is an important feature of data analysis---the choice of questions to ask about a dataset often depends on previous interactions with the same dataset. However, statistical validity is typically studied in a nonadaptive model, where all ...

This paper is concerned with stability of linear discrete time-delay systems. Note that a tighter estimation on a finite-sum term appearing in the forward difference of some Lyapunov functional leads to a less conservative delay-dependent stability ...

We prove that any non-adaptive algorithm that tests whether an unknown Boolean function f:{0,1}ⁿ→ {0,1} is a k-junta or ϵ-far from every k-junta must make Ω^˜(k^3/2) / ϵ) many queries for a wide range of parameters k and ϵ. Our result dramatically ...