The Local Embedding Problem for Hardy Spaces of Dirichlet Series

arXiv:2609.11560 · math.FA · Submitted 2026-09-10 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Mathematics. The day in mathematics: new theorems, new proofs, and what they open.

Ada: We're looking at the paper: "The Local Embedding Problem for Hardy Spaces of Dirichlet Series".

Emmy: Here's the gist of the paper. The paper addresses the local embedding problem for Hardy spaces of Dirichlet series, which is a dimension-free trace problem concerning whether the global H p-norm controls local L p-mass on the critical line Re s = 1/2.

Ada: Specifically, for every 2 < p < infinity, there exists a constant C p < infinity such that every Dirichlet polynomial P satisfies: theta in R sum from theta+1 to infinity P(1/2 + it) p dt at most C p P p H p where the constant C p is independent of the number and choice of prime variables on which P depends.

Emmy: The authors establish that this local embedding property holds exactly for p at least 2, correcting earlier conjectures that suggested integrality (p=2k) was sufficient, and noting its failure for 0 < p < 2.

Ada: The proof relies fundamentally on a single dimension-free estimate for normalized Euler products, referred to as the projected-vector estimate (Theorem 8.9).

Paper summary: Ada: So we're looking at this paper, "The Local Embedding Problem for Hardy Spaces of Dirichlet Series." Essentially, it tackles a dimension-free trace problem. The core question is whether the global H p norm controls the local L p mass on the critical line, which is where the real action happens at Re s equals one over two.

Emmy: It’s about those Dirichlet polynomials P, and for every exponent p greater than two but less than infinity, they claim there's a constant C p that works. This constant doesn't depend on how many prime variables are in the polynomial or which primes you choose for them > <ref:2609.11560#pg1>

Kolya: It sounds like it’s asserting this control: that the supremum of the sum of absolute values raised to the power p, summed over all those imaginary parts theta plus one and above, is bounded by C p times the H p norm of that Dirichlet polynomial P. That's a local embedding property.

Ada: Exactly. And this is significant because before, they only knew it for p equal to two or for even exponents like p equals two times k. They also knew it failed completely when the exponent was less than zero, specifically between zero and two > <ref:2609.11560#pg1>

Emmy: So what they establish here is that this local embedding property holds exactly for all p greater than or equal to two. That means the threshold isn't just even numbers; it’s actually the starting point of two > <ref:2609.11560#pg1>

Kolya: I wonder how they get there, though. The summary mentions a few key pieces: they rely on this single dimension-free estimate for normalized Euler products, which they call the projected-vector estimate at page seventy-eight of the paper <ref:2609.11560#pg2>. That sounds like a big technical hurdle.

Ada: Right. They use that projected-vector estimate as their fundamental building block. Their approach involves separating contributions from things that don't resonate from the part that actually resonates critically, and then controlling that resonant part using some probabilistic and harmonic analysis tools.

Emmy: And a key technical maneuver they employ is a passage to a dual exponent, q equals p over p minus one, which lands between one and two. This helps them isolate just a single resonant Euler product term through an exact frequency decomposition at page <ref:2609.11560#pg2>

Kolya: That dual exponent stuff sounds like it's where the probabilistic side comes in heavy. Does that mean the resulting constant C p is really independent of the number of primes involved? Because if it depends on those things, it’s less useful for applications.

Ada: The paper claims that C p is independent of the number and choice of prime variables on which P depends, which is a strong statement. That’s because they manage to transfer the control back from that dual exponent bound to the original Dirichlet series problem.

Emmy: It seems like this whole structure—the decomposition, the control of resonant terms via page seventy and then using proposition nine point four for discrete local sampling—is what finally gives them that local embedding inequality <ref:2609.11560#pg2>.

Kolya: So to wrap up this section, they’ve shown that even with all these technical machinery pieces like the sampling estimates and the quotient norm bounds, the final result confirms that for p greater than two, the property holds exactly > <ref:2609.11560#pg1>

Conclusion: Ada: So we've walked through how this paper tackles that local embedding problem for Hardy spaces of Dirichlet series. The authors, Bonan Chen, Xiang Fang, Feng Guo, Shengzhao Hou, Yizhou Shao, and Qi Zhou have done something important regarding the classification of these exponents.

Emmy: They started by looking at the known results at p equal to two and even integers like p equals two times k. Their work shows that this is not the complete picture; they prove that it holds exactly for all p greater than or equal to two.

Kolya: The implication for anyone listening is that if you're working in a field where you deal with these Dirichlet series, you can rely on this control for any exponent above two without worrying about needing it to be an even number. That simplifies the analysis significantly > <ref:2609.11560#pg3>

Ada: It shifts the threshold from requiring integrality to just requiring p greater than two. This is a precise classification of when this control exists for Dirichlet polynomials P and their associated Hardy spaces H p > <ref:2609.11560#pg1>

Emmy: The paper confirms that the local embedding property LE p is equivalent to having a constant C p such that the H p norm controls the L p mass for every Dirichlet polynomial. This equivalence ties it back to several classical things like conformally invariant half-plane embedding and those Gordon–Hedenmalm composition operators > <ref:2609.11560#pg2>

Kolya: From a probabilistic standpoint, the paper shows that the tail behavior governed by this estimate is well-behaved within its stated range, which is p greater than two > <ref:2609.11560#pg3>

Ada: That’s the gist of it. The main result confirms that for p greater than two, we have this specific control mechanism established through a careful decomposition of the projected vector and subsequent sampling estimates > <ref:2609.11560#pg2>

Emmy: It's a concrete result about how multiplicative structure interacts with analytic function theory on infinite polydiscs > <ref:2609.11560#pg3>

Kolya: So the paper gives us a solid boundary for this problem, showing that p equals two is the true threshold, not just the even integers above it > <ref:2609.11560#pg2>

Bonan Chen, Xiang Fang, Feng Guo, Shengzhao Hou, Yizhou Shao, Qi Zhou

math.FA

Submitted: 2026-09-10

Updated: 2026-11-16

Comments: This is an expository version and has been substantially revised with additional proof details and navigational aids. 110 pages

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Importance score: 89/100

The gist: The paper addresses the local embedding problem for Hardy spaces of Dirichlet series, which is a dimension-free trace problem concerning whether the global H p-norm controls local L p-mass on the…

Key concepts

Local Embedding Problem
This is a dimension-free question asking if the total size of a Dirichlet polynomial's values along specific vertical lines on the critical line can be bounded by its overall size in a Hardy space. It checks if global control implies local control.
Projected-Vector Estimate (Theorem 8.9)
This is a crucial estimate showing that for $p>2$, the components of a specific vector related to Euler products are controlled by the $L^q$ norm of the coefficients. This estimate is fundamental because it links complex analytic structures to simpler coefficient norms.
Critical Line Sampling
This refers to checking how well a Dirichlet series behaves when sampled at points on the line $ ext{Re } s = 1/2$. The paper shows that for $p ext{ values } ext{where } p ext{ is greater than or equal to 2}$, this sampling can be controlled by the Hardy space norm.
Resonant Term Control
The proof isolates a specific part of the Euler product called the 'resonant term' and shows it can be bounded independently of various parameters like prime dimensions or grid shifts. This control is essential for proving that only non-resonant parts affect the local embedding.

Terminology

Summary

The paper addresses the local embedding problem for Hardy spaces of Dirichlet series, which is a dimension-free trace problem concerning whether the global H p-norm controls local L p-mass on the critical line Re s = 1/2. p. 0 Specifically, for every 2 < p < infinity, there exists a constant C p < infinity such that every Dirichlet polynomial P satisfies:

theta in R sum from theta+1 to infinity P(1/2 + it) p dt at most C p |P| p H p

where the constant C p is independent of the number and choice of prime variables on which P depends. The authors establish that this local embedding property holds exactly for p at least 2, correcting earlier conjectures that suggested integrality (p=2k) was sufficient, and noting its failure for 0 < p < 2. p. 0

The proof relies fundamentally on a single dimension-free estimate for normalized Euler products, referred to as the projected-vector estimate (Theorem 8. p. 59). The methodology involves separating nonresonant contributions from the genuinely critical one, controlling the latter using probabilistic and harmonic-analytic arguments, and subsequently transferring this control back to the original Dirichlet series problem. p. 5 A key technical step involves a passage to a dual exponent q = p/(p - 1) in (1, 2), where an exact frequency decomposition isolates a single resonant Euler-product term. p. 0

The local embedding problem is shown to be equivalent to several classical formulations, including the conformally invariant half-plane embedding and the boundedness of all characteristic-zero Gordon–Hedenmalm composition operators. p. 3 The main result confirms that within these equivalent formulations, the property holds exactly for p at least 2. p. 4

The technical machinery supporting this result includes various lemmas and propositions:

  • Sampling Estimates: Lemma 7.4 provides a sampling estimate, and Proposition 7. p. 4 p. 75 establishes a frame estimate linking resonant contributions to Euler moment functionals. p. 7 Proposition 7.6 introduces a deformed frame estimate with a specific constant 0 = 1/32. p. 7

  • Decomposition: Proposition 8.5 provides an exact componentwise decomposition of the projected vector T yc into high-frequency, nonpositive, positive low, outside-shell, and resonant pieces. p. 6

  • Resonant Term Control: Proposition 8.8 yields a bound on the resonant term: Ry d L q(2) at most C p |d| q, where the constant is independent of N, the grid shift, the prime dimension, and the coefficient support.

  • Projected-Vector Estimate: Theorem 8.9 states that for every fixed p > 2, there exists a constant C p < infinity such that |T yc| L q(2(r at most y)) at most C p |c| q(Z N). The main theorem deduces the local embedding inequality from this estimate by using an exact ordered expansion of the Euler product to produce a representative in a finite-torus Hardy quotient, and then converting this bound into critical-line sampling via Proposition 9. p. 74 (Discrete local sampling). p. 1

  • Discrete Sampling: Proposition 9.4 provides a discrete local sampling estimate: for P with prime support below y=e N, there exists a constant C p such that sum from j=0 to N-1 X-1j P 1/2+it j p at most C p M y |P| p H p.

  • Quotient Norm Bound: Corollary 9.3 bounds the quotient norm: | [X j<Nc j g y, t j] L q(omega y) / N p | at most C p |c| q(Z N).

  • Martingale Estimate: Lemma 9.2 provides a first-harmonic martingale estimate: X r at most y zeta r(T yc) r L q(omega y) at most C p |c| q(Z N).

  • Critical Sampling Range: Corollary 10. p. 48 establishes the sharp critical-sampling range: for each fixed 0 < c at most 1, the equivalent properties CAS p,c and DCP p,c hold exactly for 2 at most p < infinity. p. 4

  • Final Equivalence: Proposition 10.1 confirms that the local embedding property LE p is equivalent to the existence of a constant C p such that |P| H p(i R) at most C p |P| H p for every Dirichlet polynomial P. Corollary 10.2 summarizes this, stating that assertions (i)–(vii) in Proposition 10. p. 41 hold exactly for p at least 2, while assertion (viii) holds in its stated range. p. 4

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Improved analysis of Hardy space embeddings by establishing that the local embedding property holds exactly for p ≥ 2 instead of relying on even integrality, thus providing a sharper classification for the threshold.

  2. Development of a dimension-free estimate using the projected-vector estimate (Theorem 8.9) to control nonresonant terms deterministically, which is essential for handling the missing range between even exponents.

  3. Implementation of shared-prime replacement (Proposition 3.2) to replace Euler increments with covariance-matched Gaussian first-chaos increments, allowing for a rigorous comparison between deterministic and random models by bounding the difference in expected moments (Proposition 3.6).

  4. Integration of a critical branching random walk bound to control the localized Gaussian moment estimate (Theorem 4.3), which provides a depth-dependent fractional moment estimate on an exact binary tree, crucial for deriving the final integral bounds.

  5. Application of finite cyclic harmonic analysis techniques to derive endpoint-uniform multiplier theorems (Lemma 6.1) and polynomial-bandwidth conical square function estimates (Theorem 6.4), which provide uniform control over the prime structure independent of the total size N = 2H.

  6. Construction of a common-shift moment bound (Proposition 5.3) that controls the expected fractional moment of the total coefficient energy by the conical square function, effectively linking localized Euler moments to global estimates without requiring independence between function levels.

  7. Utilization of deterministic tent allocation (Lemma 5.2) to partition nonzero coefficient cells into disjoint tents whose total energy is controlled by a uniform bound on the conical square function, leading directly to the final trace estimate (Proposition 6.6).

Related papers