Weighted decoupling with lower-dimensional frequency localization

Technical Summary: Weighted Decoupling with Lower-Dimensional Frequency Localization

Problem Statement
This paper addresses the problem of establishing weighted $L^2$ and refined $L^p$ decoupling estimates for functions whose Fourier transforms are supported in a small neighborhood ($R^{-1}$-neighborhood) of a compact strictly convex $C^2$ hypersurface $S \subset \mathbb{R}^d$ (such as the unit sphere or truncated paraboloid). The central novelty lies in incorporating an additional lower-dimensional frequency localization property. Specifically, the functions are assumed to be sums of wave packets whose long directions are concentrated near an $m$-dimensional subspace $V \subset \mathbb{R}^d$ (referred to as having $(r, m)$-concentrated frequencies).

The work is motivated by the Falconer distance set problem, which seeks to determine the Hausdorff dimension threshold for a set $E \subset \mathbb{R}^d$ such that its distance set $\Delta(E)$ has positive Lebesgue measure. Recent progress on this problem relies on decomposing fractal measures into "good" and "bad" parts, where the "good" part requires decoupling estimates for functions with concentrated frequencies. The paper aims to improve the bounds on these estimates, particularly for the case $d=3$, to potentially sharpen the known dimension thresholds.

Methodology
The proof strategy follows the standard broad–narrow analysis framework, adapted for weighted estimates and lower-dimensional concentration:

1. Broad Case:

   • The domain is partitioned into "broad" cubes where the wave packets exhibit transversality in multiple directions.
   • The authors combine refined decoupling estimates with multilinear Kakeya estimates (Bennett-Carbery-Tao) to derive a multilinear weighted $L^2$ estimate.
   • This approach avoids technical arguments found in previous works (e.g., Du-Zhang) regarding the interchange of products and integrals by utilizing a formulation inspired by Demeter.

2. Narrow Case:

   • The domain is partitioned into "narrow" cubes where wave packets are concentrated near a lower-dimensional subspace.
   • The proof employs an induction on scales argument. The scale is reduced from $R$ to $R/K^2$ via parabolic scaling.
   • A key technical step involves tracking the density parameters of the weight set $Y$ under the affine transformations induced by parabolic scaling. The authors establish how the density parameters $\gamma_{\alpha, R}$ and $\gamma_{\beta, R}$ transform, ensuring the induction closes with the necessary gain in the error term.

3. Weighted Framework:

   • The paper introduces specific density parameters $\gamma_{\alpha, R}(Y)$ and a composite quantity $A_{d,m,\alpha,\beta}(R, Y)$ to quantify the distribution of the weight set $Y$ at different scales ($R$ and $R^{1/2}$).
   • The estimates are derived by interpolating between weighted $L^2$ estimates (derived from refined decoupling) and standard refined decoupling estimates.

Key Contributions and Results

• Theorem 1.1 (Weighted $L^2$ Decoupling): The authors prove a weighted $L^2$ estimate for functions with $(R^{-1/2}, m)$-concentrated frequencies. The bound depends on the density parameters of the weight set $Y$ at scales $R$ and $R^{1/2}$.

  • Significance: This result recovers the fractal $L^2$ restriction estimate of Du and Zhang but with a sharper dependence on the density of the weight. Specifically, the product of density factors $\gamma_{\alpha, R^{1/2}}(Y)\gamma_{\alpha, R}(Y)$ replaces the squared factor $\gamma_{\alpha, R}(Y)^2$ found in earlier works, providing a significant improvement when the density at scale $R^{1/2}$ is much smaller than at scale $R$.

• Theorem 1.5 (Schrödinger Maximal Function): As a corollary, the paper derives a weighted estimate for the Schrödinger maximal function (associated with the paraboloid). This sharpens the result of Du and Zhang [DZ19], offering a more precise bound when the weight distribution is non-uniform across scales.

• Theorem 1.7 (Weighted Refined Decoupling): The paper establishes a weighted refined decoupling estimate for $L^p$ norms ($2 \le p \le p_m$).

  • Improvement: Under the assumption that the weight is $\alpha$-dimensional at every scale, this result improves upon the bound in Du, Ou, Ren, and Zhang [DORZ23b] when $\alpha < m$. The improvement stems from the refined handling of the density parameters at different scales.

• Proposition 1.6 & Sharpness: The authors construct examples demonstrating that their weighted refined decoupling estimates are sharp for certain configurations of wave packets and weight sets.

• Theorem 1.8 (Falconer Distance Set in $\mathbb{R}^3$): Applying the new decoupling estimates to the Falconer problem in dimension $d=3$, the authors prove that for a compact set $E \subset \mathbb{R}^3$, the pinned distance set $\Delta_x(E)$ has positive Lebesgue measure for some $x \in E$ whenever $\dim_H(E) > \frac{\sqrt{241}-5}{6} \approx 1.754$.

  • Context: This represents a modest improvement over the threshold of $\approx 1.757$ obtained in [DORZ23b] using the same decomposition framework. However, the authors explicitly note that this result does not recover the stronger threshold of $12/7 \approx 1.714$ obtained in [DORZ23a], indicating that further improvements in decoupling estimates are required to close the gap within the current framework.

Significance and Claims
The paper claims to provide a partial progress toward resolving the Falconer distance set conjecture in $\mathbb{R}^3$ within the specific framework of [DORZ23b]. By refining the weighted decoupling estimates to account for lower-dimensional frequency concentration and multi-scale density properties, the authors achieve a sharper bound on the Schrödinger maximal function and a slightly improved dimension threshold for the Falconer problem.

The authors are modest in their claims regarding the Falconer problem, acknowledging that while their method improves the threshold from [DORZ23b], it falls short of the best-known result in [DORZ23a]. The primary contribution is the technical advancement in the decoupling machinery itself, specifically the handling of weights with varying densities at different scales and the integration of lower-dimensional frequency concentration, which may have applications beyond the immediate scope of the Falconer problem.