Pointwise convergence of bilinear ergodic averages for commuting transformations

arXiv:2610.26388 · math.DS · Submitted 2026-10-19 · Read on arXiv

math.DS

Submitted: 2026-10-19

Updated: 2026-10-24

Comments: Revised proof and updated references. Added a note explaining that the main theorem is the same as the result previously formalized by van Doorn, Durcik, Roos, Slavíková and Thiele (Palomar, PALOMAR-2026-10-18-000026 v1). At present, we do not plan to submit this manuscript for journal publication

License: http://creativecommons.org/licenses/by-nc-nd/4.0/

The gist: Let (X, B,μ) be a probability space and let S,Q:X to X be commuting probability-preserving transformations.

Terminology

Abstract

Let (X, B,μ) be a probability space and let S,Q:X to X be commuting probability-preserving transformations. Let 1 at most p,q at most infinity satisfy 1/p+1/q at most1. We prove that, for every fixed F in L p(μ) and G in L q(μ), the averages 1N sum from n=1 to N F(S nx)G(Q nx) converge almost everywhere as N to infinity. Our proof introduces a new analytic approach that relies on neither the circle method nor characteristic-factor theory. Exploiting the triangular structure arising from commuting transformations, we combine matrix-energy estimates with Gaussian heat-flow arguments to establish pointwise variation inequalities for suitable smooth averages. We then use even and odd Gaussian kernels to recover the convergence of the original averages.

Sources

Related papers