On Lipschitz bijections between Boolean functions

From MaRDI portal
(Redirected from Publication:4635511)



Abstract: For two functions f,g:0,1no0,1 a mapping psi:0,1no0,1n is said to be a ftog if it is a bijection and f(z)=g(psi(z)) for every zin0,1n. In this paper we study Lipschitz mappings between boolean functions. Our first result gives a construction of a C-Lipschitz mapping from the sfMajority function to the sfDictator function for some universal constant C. On the other hand, there is no n/2-Lipschitz mapping in the other direction, namely from the sfDictator function to the sfMajority function. This answers an open problem posed by Daniel Varga in the paper of Benjamini et al. (FOCS 2014). We also show a mapping from sfDictator to sfXOR that is 3-local, 2-Lipschitz, and its inverse is O(log(n))-Lipschitz, where by L-local mapping we mean that each of its output bits depends on at most L input bits. Next, we consider the problem of finding functions such that any mapping between them must have large emph{average stretch}, where the average stretch of a mapping phi is defined as sfavgStretch(phi)=mathbbEx,i[dist(phi(x),phi(x+ei)]. We show that any mapping phi from sfXOR to sfMajority must satisfy sfavgStretch(phi)geqOmega(sqrtn). In some sense, this gives a "function analogue" to the question of Benjamini et al. (FOCS 2014), who asked whether there exists a set Asubset0,1n of density 0.5 such that any bijection from 0,1n1 to A has large average stretch. Finally, we show that for a random balanced function f:0,1no0,1n with high probability there is a mapping phi from sfDictator to f such that both phi and phi1 have constant average stretch. In particular, this implies that one cannot obtain lower bounds on average stretch by taking uniformly random functions.











This page was built for publication: On Lipschitz bijections between Boolean functions

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4635511)