Cancellation for the multilinear Hilbert transform
Classical Analysis and ODEs
2015-06-01 v3 Combinatorics
Abstract
For any natural number k, consider the k-linear Hilbert transform Hk(f1,…,fk)(x):=p.v.∫Rf1(x+t)…fk(x+kt) tdt for test functions f1,…,fk:R→C. It is conjectured that Hk maps Lp1(R)×⋯×Lpk(R)→Lp(R) whenever 1<p1,…,pk,p<∞ and p1=p11+⋯+pk1. This is proven for k=1,2, but remains open for larger k. In this paper, we consider the truncated operators Hk,r,R(f1,…,fk)(x):=∫r≤∣t∣≤Rf1(x+t)…fk(x+kt) tdt for R>r>0. The above conjecture is equivalent to the uniform boundedness of ∥Hk,r,R∥Lp1(R)×⋯×Lpk(R)→Lp(R) in r,R, whereas the Minkowski and H\"older inequalities give the trivial upper bound of 2logrR for this quantity. By using the arithmetic regularity and counting lemmas of Green and the author, we improve the trivial upper bound on ∥Hk,r,R∥Lp1(R)×⋯×Lpk(R)→Lp(R) slightly to o(logrR) in the limit rR→∞ for any admissible choice of k and p1,…,pk,p. This establishes some cancellation in the k-linear Hilbert transform Hk, but not enough to establish its boundedness in Lp spaces.
Cite
@article{arxiv.1505.06479,
title = {Cancellation for the multilinear Hilbert transform},
author = {Terence Tao},
journal= {arXiv preprint arXiv:1505.06479},
year = {2015}
}
Comments
17 pages, no figures. An error pointed out to the author by Pavel Zorin-Kranich has been corrected