theoretical-computer-science / Analysis of Boolean functions; polynomial threshold functions
The asymptotic Gotsman-Linial conjecture on average sensitivity of polynomial threshold functions
Theorem 1.1: for n≥1, 1≤d≤n and a real multilinear polynomial p of degree at most d, f=sgn(p) with sgn(0)=1 has I(f)≤8dn; the constant is absolute and d may grow with n. Symmetric examples show the order dn is sharp for d≤n. It removes the polylogarithmic loss in Kane's bound. It makes no claim about the exact extremizers of the original Gotsman-Linial proposal (already refuted by Chapman and by Kim-Maldonado-Wellens) and does not address non-uniform distributions.
Theorem 1.1: for n≥1, 1≤d≤n and a real multilinear polynomial p of degree at most d, f=sgn(p) with sgn(0)=1 has I(f)≤8dn; the constant is absolute and d may grow with n. Symmetric examples show the order dn is sharp for d≤n. It removes the polylogarithmic loss in Kane's bound. It makes no claim about the exact extremizers of the original Gotsman-L…
Theorem 1.1: for n≥1, 1≤d≤n and a real multilinear polynomial p of degree at most d, f=sgn(p) with sgn(0)=1 has I(f)≤8dn; the constant is absolute and d may grow with n. Symmetric examples show the order dn is sharp for d≤n. It removes the polylogarithmic loss in Kane's bound. It makes no claim about the exact extremizers of the original Gotsman-Linial proposal (already refuted by Chapman and by Kim-Maldonado-Wellens) and does not address non-uniform distributions.
Theorem 1.1: for n≥1, 1≤d≤n and a real multilinear polynomial p of degree at most d, f=sgn(p) with sgn(0)=1 has I(f)≤8dn; the constant is absolute and d may grow with n. Symmetric examples show the order dn is sharp for d≤n. It removes the polylogarithmic loss in Kane's bound. It makes no claim about the exact extremizers of the original Gotsman-Linial proposal (already refuted by Chapman and by Kim-Maldonado-Wellens) and does not address non-uniform distributions.