Skip to yearly menu bar Skip to main content


Poster

Sharp Rates in Dependent Learning Theory: Avoiding Sample Size Deflation for the Square Loss

Ingvar Ziemann · Stephen Tu · George J. Pappas · Nikolai Matni


Abstract: In this work, we study statistical learning with dependent data and square loss in a hypothesis class with tail decay in Orlicz space: $\mathscr{F}\subset L_{\Psi_p}$. Our inquiry is motivated by the search for a sharp noise interaction term, or variance proxy, in learning with dependent (e.g. $\beta$-mixing) data. Typical non-asymptotic results exhibit variance proxies that are deflated *multiplicatively* in the mixing time of the underlying covariates process. We show that whenever the topologies of $L^2$ and $\Psi_p$ are comparable on our hypothesis class $\mathscr{F}$, the empirical risk minimizer achieves a rate that only depends on the complexity of the class and second order statistics in its leading term. We refer to this as a *near mixing-free rate*, since direct dependence on mixing is relegated to an additive higher order term. Our approach, reliant on mixed tail generic chaining, allows us to obtain sharp, instance-optimal rates. Examples that satisfy our framework include for instance sub-Gaussian linear regression and bounded smoothness classes.

Live content is unavailable. Log in and register to view live content