An optimal lower bound for monotonicity testing over hypergrids

From MaRDI portal



Abstract: For positive integers n,d, consider the hypergrid [n]d with the coordinate-wise product partial ordering denoted by prec. A function f:[n]dmapstomathbbN is monotone if forallxprecy, f(x)leqf(y). A function f is eps-far from monotone if at least an eps-fraction of values must be changed to make f monotone. Given a parameter eps, a emph{monotonicity tester} must distinguish with high probability a monotone function from one that is eps-far. We prove that any (adaptive, two-sided) monotonicity tester for functions f:[n]dmapstomathbbN must make Omega(eps−1dlogn−eps−1logeps−1) queries. Recent upper bounds show the existence of O(eps−1dlogn) query monotonicity testers for hypergrids. This closes the question of monotonicity testing for hypergrids over arbitrary ranges. The previous best lower bound for general hypergrids was a non-adaptive bound of Omega(dlogn).











This page was built for publication: An optimal lower bound for monotonicity testing over hypergrids

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2851875)