Multilinear Bochner-Riesz Operator

• Multilinear Bochner–Riesz operator is a higher-order Fourier multiplier that generalizes classical Bochner–Riesz means to regularize joint spectral projections with a smoothness parameter.
• It employs dyadic radial decompositions, square-function techniques, and real interpolation to establish sharp L^p boundedness thresholds and analyze endpoint phenomena.
• Recent advances link this operator to multilinear restriction theory and maximal variants, paving the way for improved estimates and new directions in harmonic analysis.

A multilinear Bochner–Riesz operator is a higher-order Fourier multiplier operator central to modern harmonic analysis, generalizing the classical Bochner–Riesz means to multilinear (and especially bilinear and trilinear) settings. Such operators regularize the joint spectral projections associated with sums of squares of multiple frequencies, imposing a smoothness parameter to control singularities and oscillatory behavior near critical geometric surfaces (typically spheres) in higher-dimensional frequency space. The study encompasses $L^p$ boundedness, maximal and square-function extensions, endpoint phenomena, and deep connections to multilinear restriction theory and oscillatory integral estimates.

1. Definition and Notation

For integers $n \geq 1$ (dimension) and $m \geq 2$ (degree of multilinearity), the $m$-linear Bochner–Riesz operator of order $\delta \geq 0$ is defined via the Fourier multiplier

$$M^\delta(\xi_1, \dots, \xi_m) = \left(1 - |\xi_1|^2 - \cdots - |\xi_m|^2\right)_+^\delta,$$

where $(\cdot)_+ = \max\{\cdot, 0\}$. For Schwartz functions $f_1, \dots, f_m$, the associated operator acts as

$$B^\delta(f_1, \dots, f_m)(x) = \int_{\mathbb{R}^{nm}} M^\delta(\xi_1, \dots, \xi_m) \prod_{j=1}^m \widehat{f_j}(\xi_j) e^{2\pi i x \cdot (\xi_1 + \cdots + \xi_m)} \, d\xi_1 \cdots d\xi_m.$$

The bilinear ($m=2$) and trilinear ($n \geq 1$0) cases are of particular prominence. In many contexts, maximal variants and associated square functions, such as

$n \geq 1$1

and

$n \geq 1$2

are studied to probe convergence and regularity phenomena (Shuin, 2024, He et al., 2024).

2. Boundedness Results and Critical Exponents

Classical interest is in $n \geq 1$3 boundedness, where $n \geq 1$4. Boundedness thresholds for $n \geq 1$5 are governed by the geometric singularity of the multiplier and the interplay of $n \geq 1$6.

For the bilinear case, fundamental theorems establish:

• $n \geq 1$7 is bounded if and only if $n \geq 1$8 for $n \geq 1$9 (Bernicot et al., 2012).
• More general $m \geq 2$0 bounds hold provided $m \geq 2$1, with an explicit function $m \geq 2$2 determined by the position in the index simplex.
• For very smooth case, i.e., $m \geq 2$3, trivial kernel estimates yield boundedness for all indices.
• Similar threshold phenomena extend to higher-order multilinear operators; for $m \geq 2$4-linear maximal means, one has sharp bounds for all

$m \geq 2$5

where $m \geq 2$6 (Shuin, 2024, He et al., 2024).

Recent advances for the bilinear Bochner–Riesz operator include scale-invariant sharp bounds and improvements at non-endpoint exponents using linear square function techniques and almost-orthogonality decompositions (Jeong et al., 2017). For multilinear maximal and square-function variants, induction on $m \geq 2$7 with dyadic decompositions and real interpolation yields sharp results including regimes $m \geq 2$8 inaccessible to previous Calderón–Zygmund theory (Shuin, 2024, He et al., 2024).

A tabulation of critical thresholds for the $m \geq 2$9-linear maximal operator:

Multilinearity $m$0	Threshold for $m$1	Notable features
$m$2 (linear)	$m$3	Classical Bochner–Riesz
$m$4 (bilinear)	$m$5	New endpoint sharpness
$m$6	$m$7	Additivity via induction

3. Methodologies

The analysis of multilinear Bochner–Riesz operators relies on an overview of advanced techniques:

• Dyadic Radial (Spherical) Decomposition: Decomposing the multiplier into frequency shells via

$m$8

localizes analysis to annuli around the singularity (Bernicot et al., 2012).

• Fourier Series and Restriction–Extension Arguments: Fourier expansion in radial variables or reduction to restriction–extension operators, relating the problem to spherical means and associated $m$9 estimates.
• Square Functions: The boundedness of multilinear Bochner–Riesz is closely connected to square-function estimates (discretized and continuous). Recent work exploits known and conjectured sharp bounds for Stein’s and Carbery’s square functions (Jeong et al., 2017, He et al., 2024).
• Induction and Real Interpolation: Inductive arguments on multilinearity, combined with almost-orthogonality in dyadic scales and interpolation between critical exponents, fill in ranges for nontrivial $\delta \geq 0$0 mapping (Shuin, 2024, He et al., 2024).
• Wavelet Decomposition: For maximal variants, off-diagonal wavelet decompositions allow geometric decay estimates for dyadic pieces, which sum to global operator bounds (He, 2016).

4. Model Operators, Simultaneous Saturation, and Multilinear Restriction Linkage

Multilinear Bochner–Riesz analysis is tied to restriction theory by the geometric structure of the singular set $\delta \geq 0$1. The modern perspective employs "model operators" formulated as products of oscillatory-integral operators localized to directions, leading to a principle of simultaneous saturation.

• The principle of simultaneous saturation furnishes a framework to control the size of level sets where multiple oscillatory integrals are large, and leads to sharp $\delta \geq 0$2 estimates for multilinear models (Tacy, 30 Jan 2025).
• In classical restriction theory, optimality depends on strict transversality of phase gradients; simultaneous saturation relaxes this by quantifying allowable degeneracy, yielding robust multilinear analogues relevant for the Bochner–Riesz conjecture.

This connection suggests that the multilinear Bochner–Riesz problem can benefit from techniques and conjectures from restriction theory, specifically the Bennett–Carbery–Tao multilinear restriction framework.

5. Endpoint Phenomena, Sharpness, and Open Problems

A salient outcome is the identification of sharp or near-sharp indices for boundedness and convergence:

• For the bilinear operator, $\delta \geq 0$3 boundedness holds for all $\delta \geq 0$4, sharp in the sense that no such bound holds at $\delta \geq 0$5 (Bernicot et al., 2012).
• For maximal Bochner–Riesz (bilinear), sharpness is seen at $\delta \geq 0$6; below this threshold, boundedness fails (He, 2016).
• In the $\delta \geq 0$7-linear case, new threshold phenomena emerge for $\delta \geq 0$8 when some exponents $\delta \geq 0$9; in many low-$$M^\delta(\xi_1, \dots, \xi_m) = \left(1 - |\xi_1|^2 - \cdots - |\xi_m|^2\right)_+^\delta,$$0 regimes, these are the first positive results (He et al., 2024). Explicit wave-packet counterexamples demonstrate that the gap between necessary and sufficient conditions is often at most 1.
• Endpoint $$M^\delta(\xi_1, \dots, \xi_m) = \left(1 - |\xi_1|^2 - \cdots - |\xi_m|^2\right)_+^\delta,$$1 ("disc multiplier") boundedness remains open except for special indices; full characterization in higher dimensions is unresolved (Jeong et al., 2017).

Open questions center on closing remaining gaps at endpoint exponents and in the multidimensional index simplex, extending sharpness for more general weights and vector-valued variants, and understanding whether minimal smoothness or transversality can be further relaxed—especially in connection with restriction conjectures.

6. Weighted Estimates and Convergence Properties

Advances include robust weighted $$M^\delta(\xi_1, \dots, \xi_m) = \left(1 - |\xi_1|^2 - \cdots - |\xi_m|^2\right)_+^\delta,$$2 theory for multilinear Bochner–Riesz means and maximal operators. For weights $$M^\delta(\xi_1, \dots, \xi_m) = \left(1 - |\xi_1|^2 - \cdots - |\xi_m|^2\right)_+^\delta,$$3 in the Muckenhoupt classes $$M^\delta(\xi_1, \dots, \xi_m) = \left(1 - |\xi_1|^2 - \cdots - |\xi_m|^2\right)_+^\delta,$$4,

$$M^\delta(\xi_1, \dots, \xi_m) = \left(1 - |\xi_1|^2 - \cdots - |\xi_m|^2\right)_+^\delta,$$5

with the threshold for $$M^\delta(\xi_1, \dots, \xi_m) = \left(1 - |\xi_1|^2 - \cdots - |\xi_m|^2\right)_+^\delta,$$6 coinciding with the unweighted case (He et al., 2024). The proofs utilize vector-valued maximal function inequalities (e.g., Fefferman–Stein), square-function generalizations, and subtle reproducing kernel decompositions.

Pointwise convergence of multilinear Bochner–Riesz means to products of initial functions is established for exponents and $$M^\delta(\xi_1, \dots, \xi_m) = \left(1 - |\xi_1|^2 - \cdots - |\xi_m|^2\right)_+^\delta,$$7 above the sharp critical index $$M^\delta(\xi_1, \dots, \xi_m) = \left(1 - |\xi_1|^2 - \cdots - |\xi_m|^2\right)_+^\delta,$$8; at the threshold, only weak-type or conditional convergence is generally available (He et al., 2024).

7. Outlook and Future Directions

Current research highlights include:

• Systematic reduction of the multilinear Bochner–Riesz problem to square-function technology, leveraging almost-orthogonality, and induction on multilinearity.
• The role of simultaneous saturation and nested transversality conditions, especially for the adaptation of model oscillatory integral operators to actual Bochner–Riesz multipliers (Tacy, 30 Jan 2025).
• Multilinear maximal and square-function variants are fundamentally additive in their critical exponents, paralleling classical Calderón–Zygmund multilinear theory but requiring deeper geometric and combinatorial decomposition.
• The extension to non-integer order, variable coefficients, and geometric measure settings, as well as connections to decoupling and Kakeya-type phenomena.

The multilinear Bochner–Riesz operator continues to stimulate development in harmonic analysis, interacting with restriction theory, oscillatory integral bounds, and weighted norm inequalities, with open questions at the interface of analysis, combinatorics, and geometric measure theory.

---

References

• (Bernicot et al., 2012) The bilinear Bochner-Riesz problem
• (Jeong et al., 2017) Improved bound for the bilinear Bochner-Riesz operator
• (Shuin, 2024) $$M^\delta(\xi_1, \dots, \xi_m) = \left(1 - |\xi_1|^2 - \cdots - |\xi_m|^2\right)_+^\delta,$$9 estimates for multilinear maximal Bochner--Riesz means and square function
• (He et al., 2024) On pointwise convergence of multilinear Bochner-Riesz means
• (Tacy, 30 Jan 2025) The principle of simultaneous saturation: applications to multilinear models for the restriction and Bochner-Riesz problem
• (He, 2016) On Bilinear Maximal Bochner-Riesz Operators