ℓ p ​ ( ℤ n ) \ell^{p}({\mathbb{Z}}^{n}) -estimate for long r r -variational seminorm of discrete Birch-Magyar averages

Ankit Bhojak, Siddhartha Samanta, Saurabh Shrivastava · 2026

Read ℓ p ​ ( ℤ n ) \ell^{p}({\mathbb{Z}}^{n}) -estimate for long r r -variational seminorm of discrete Birch-Magyar averages by Ankit Bhojak, Siddhartha Samanta, Saurabh Shrivastava (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-00654.txt.

Introduction

Ankit Bhojak and two co-authors prove bounds in a 2026 math paper on averages over whole-number points. The averages at stake sum values over special curved sets of points. One family is named for Birch and Magyar, and one is of Hardy-Littlewood type. The authors track how much these averages can swing as the scale of the set grows. That swing is summed by a gauge called a variational seminorm.

Their bounds say the swing stays in check for a wide range of gauges. The range ties the strength of the norm to the swing gauge in a sharp way. The authors also draw a use in ergodic theory. That field studies long-run runs of systems that keep some total fixed. Bounds like these help show when time averages match space averages. The paper adds firm ground to a hard branch of analysis.

Catalog record

CREATORAnkit Bhojak, Siddhartha Samanta, Saurabh Shrivastava
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE43 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2601-00654

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

ℓ p ​ ( ℤ n ) \ell^{p}({\mathbb{Z}}^{n}) -estimate for long r r -variational seminorm of discrete Birch-Magyar averages Ankit Bhojak, Siddhartha Samanta, Saurabh Shrivastava Abstract . We prove ℓ p ​ ( ℤ n ) − \ell^{p}({\mathbb{Z}}^{n})- estimates for long r r -variational seminorm of two families of averages: discrete Birch-Magyar averages, for r > m ​ a ​ x ​ { p , p ′ } r>max\{p,p^{\prime}\} with p > 2 ​ c ℜ − 2 2 ​ c ℜ − 3 p>\frac{2c_{\mathfrak{R}}-2}{2c_{\mathfrak{R}}-3} and discrete Hardy-Littlewood type averages over certain algebraic varieties, for r > m ​ a ​ x ​ { p , p ′ } r>max\{p,p^{\prime}\} with p > 1 p>1 . Further, we discuss an application of these results in ergodic theory. 2020 Mathematics Subject Classification Primary 11P55, 11D72, Secondary 42B25, 37A46 1. Introduction Let D\subset{\mathbb{N}} , and f:{\mathbb{Z}}^{n}\rightarrow{\mathbb{C}} be a function in \ell^{p}({\mathbb{Z}}^{n}) with 1\leq p\leq\infty . For a family of operators \mathcal{A}:=\{A_{\lambda}:\lambda\in D\} acting on f and r\in[1,\infty] , we define the r -variations corresponding to the family \mathcal{A}f=\{A_{\lambda}f:\lambda\in D\} by (1.1) \displaystyle V_{r}(\mathcal{A}f)(x):=\begin{cases}\sup\limits_{L\in{\mathbb{N}}}\sup\limits_{\lambda_{1}<\dots<\lambda_{L}}\left(\sum\limits_{l=1}^{L-1}|A_{\lambda_{l+1}}f(x)-A_{\lambda_{l}}f(x)|^{r}\right)^{\frac{1}{r}},&0<r<\infty,\\ \sup\limits_{\lambda_{0}<\lambda_{1}}\left|A_{\lambda_{1}}f(x)-A_{\lambda_{0}}f(x)\right|,&r=\infty,\end{cases} where the former supremum is taken over all finite increasing sequences \{\lambda_{1},\dots,\lambda_{L}\} belonging to D . The study of r -variation was initiated in the work of Lépingle [ 13 ] . Bourgain’s seminal work [ 5 ] on variational estimates for ergodic averages introduced a new direction of research in harmonic analysis. Since then, extensive studies of r -variation for various classes of discrete and continuous averages have been conducted. We refer to [ 4 , 19 , 18 , 20 , 11 ] and references therein for more details. In partiuclar, Jones, Seeger, and Wright [ 11 ] studied L^{p}- estimates of the r -variation for continuous spherical averages. To the best of our knowledge, the study of variational estimates for the class of discrete averages over certain surfaces, which includes the sphere, are not known. The purpose of this article is to initiate the study of \ell^{p}(\mathbb{Z}^{n}) -boundedness of r -variations of the discrete Birch-Magyar averages and discrete Hardy-Littlewood type averages over certain algebraic varieties. 1.1. Birch-Magyar averages The framework for the Birch-Magyar averages was developed in the work of Birch [ 3 ] and Magyar [ 17 ] . Let us first briefly introduce the notion of Birch-Magyar averages. Recall that a homogeneous polynomial with integral coefficients is called an integral form. Let \mathfrak{R}(x) be an integral form of degree d>1 in n variables. The Birch rank of \mathfrak{R} , denoted by \mathcal{B}(\mathfrak{R}), is defined by \mathcal{B}(\mathfrak{R})=n-\text{dim}(\mathcal{N}(\mathfrak{R})), where \mathcal{N}(\mathfrak{R})=\{z\in\mathbb{C}^{n}:\partial_{z_{1}}\mathfrak{R}(z)=\partial_{z_{2}}\mathfrak{R}(z)=\cdots=\partial_{z_{n}}\mathfrak{R}(z)=0\}. We use the following standard notation c_{\mathfrak{R}}=\frac{\mathcal{B}(\mathfrak{R})}{(d-1)2^{d-1}}~~\text{and}~~\eta_{\mathfrak{R}}=\frac{1}{6d}\left(\frac{c_{\mathfrak{R}}}{2}-1\right). Note that d\eta_{\mathfrak{R}}<c_{\mathfrak{R}}-2 . For a given compactly supported smooth function \phi with 0\leq\phi\leq 1 , an integral form \mathfrak{R}(x) of degree d>1 in n variables is called \phi- regular if there exists a non-singular solution of \mathfrak{R}(x)=1 in the support of \phi and Birch rank of \mathfrak{R}(x) is greater than (d-1)2^{d} . We note that c_{\mathfrak{R}}>2 if \mathfrak{R}(x) is \phi- regular. Consider the counting function r_{\mathfrak{R},\phi}(\lambda)=\sum\limits_{\mathfrak{R}(x)=\lambda}\phi\left(\frac{x}{\lambda^{\frac{1}{d}}}\right) . It is known, see [ 3 , 17 ] , that for a \phi- regular form \mathfrak{R}(x), there is an infinite arithmetic progression \Lambda of regular values \lambda\in\Lambda satisfying r_{\mathfrak{R},\phi}(\lambda)\approx\lambda^{\frac{n}{d}-1} . Henceforth, for a \phi- regular form \mathfrak{R}(x) , let \Lambda denote such an infinite arithmetic progression. We refer the readers to [ 3 ] , [ 17 ] , and [ 6 ] for more details on \phi- regular forms. For a \phi- regular form \mathfrak{R}(x) of degree d>1 in n variables and \lambda\in\Lambda , the Birch-Magyar average of f is defined by (1.2) \displaystyle A_{\lambda}^{\mathfrak{R}}f(x)=\frac{1}{r_{\mathfrak{R},\phi}(\lambda)}\sum_{y:\mathfrak{R}(y)=\lambda}\phi\left(\frac{y}{\lambda^{\frac{1}{d}}}\right)f(x-y). Note that the definition above encompasses the k- spherical averages studied in [ 16 ] , [ 15 ] , [ 10 ] , [ 9 ] , and [ 1 ] . Recall that a sequence \{\lambda_{l}\}_{l=1}^{\infty} is called lacunary if there exists c>1 such that \frac{\lambda_{l+1}}{\lambda_{l}}\geq c, for all l\geq 1. For a lacunary sequence \mathbb{L}=\{\lambda_{l}\}_{l=1}^{\infty}\subset\Lambda , the lacunary maximal function associated to the Birch-Magyar averages is defined as A_{lac}^{\mathfrak{R}}f(x)=\sup\limits_{\lambda\in\mathbb{L}}|A_{\lambda}^{\mathfrak{R}}f(x)|. Cook and Hughes [ 6 ] prove that the lacunary maximal operator A_{lac}^{\mathfrak{R}} is bounded on \ell^{p}(\mathbb{Z}^{n}) for p>\frac{2c_{\mathfrak{R}}-2}{2c_{\mathfrak{R}}-3} extending the result of Kesler, Lacey and Mena [ 12 ] in the spherical case. Moreover, building on a counterexample of Zienkiewicz, they showed that there exists a lacunary sequence \{\lambda_{l}\}_{l=1}^{\infty} such that the corresponding maximal operator A_{lac}^{\mathfrak{R}} is not bounded on \ell^{p}(\mathbb{Z}^{n}) for the range 1<p<\frac{n}{n-1} . However, it remains unknown that whether the maximal operator A_{lac}^{\mathfrak{R}} is bounded on \ell^{p}({\mathbb{Z}}^{n}) in the range \frac{n}{n-1}\leq p\leq\frac{2c_{\mathfrak{R}}-2}{2c_{\mathfrak{R}}-3} . We refer the readers to [ 12 , 8 , 2 ] for discrete maximal functions over sparser sequences, for which the boundedness does hold for p>1 . Let \mathcal{A}_{\mathbb{L}}^{{\mathfrak{R}}}=\{A_{\lambda}^{{\mathfrak{R}}}:\;\lambda\in\mathbb{L}\} be the family of Birch-Magyar averages over lacunary sequence \mathbb{L}\subset\>\Lambda . We define the long r- variations associated to the family \mathcal{A}_{\mathbb{L}}^{{\mathfrak{R}}} as follows. V_{r}(\mathcal{A}_{\mathbb{L}}^{\mathfrak{R}}f)(x):=V_{r}(\{A_{\lambda}^{\mathfrak{R}}f,\;\lambda\in\mathbb{L}\})(x). We note that |A_{lac}^{\mathfrak{R}}f(x)|\leq|A_{\lambda_{0}}^{{\mathfrak{R}}}f(x)|+2V_{r}(\mathcal{A}_{\mathbb{L}}^{\mathfrak{R}}f)(x), holds for any \lambda_{0}\in{\mathbb{N}} and 1\leq r<\infty . Thus, the \ell^{p}- bounds for the long r- variations imply the corresponding bounds for the lacunary maximal operator A_{lac}^{\mathfrak{R}} . Our first main result concerns the \ell^{p}- bounds for the long r- variations associated to the Birch-Magyar averages. We have the following. Theorem 1.1 . Let \mathfrak{R}(x) be a \phi- regular form in n variables of degree d>1 and \mathbb{L}\subset\Lambda be a lacunary sequence. Then for r>max\{p,p^{\prime}\} with p>\frac{2c_{\mathfrak{R}}-2}{2c_{\mathfrak{R}}-3} , we have (1.3) \|V_{r}(\mathcal{A}_{\mathbb{L}}^{\mathfrak{R}}f)\|_{\ell^{p}({\mathbb{Z}}^{n})}\lesssim\|f\|_{\ell^{p}({\mathbb{Z}}^{n})}, where p^{\prime} is the Hölder conjugate of p given by \frac{1}{p}+\frac{1}{p^{\prime}}=1 . The range of r in the above theorem is sharp for p=2 , see for instance [ 11 ] . In view of the pointwise inequality V_{\infty}(\mathcal{A}_{\mathbb{L}}^{{\mathfrak{R}}}f)(x)\leq 2A_{lac}^{\mathfrak{R}}f(x) , the \ell^{p}- boundedness of A_{lac}^{\mathfrak{R}} ( [ 6 , Theorem 1] ) for p>\frac{2c_{\mathfrak{R}}-2}{2c_{\mathfrak{R}}-3} and interpolation, it is enough to prove Theorem 1.1 for p=2 . Moreover, to prove the inequality ( 1.3 ) for p=2 , we will rely on the jump inequalities corresponding to the family \mathcal{A}_{\mathbb{L}}^{{\mathfrak{R}}} . In that regard, we define the \lambda -jump corresponding to the family \mathcal{A}_{\mathbb{L}}^{{\mathfrak{R}}} , denoted by \mathcal{J}_{\lambda}(\mathcal{A}_{\mathbb{L}}^{{\mathfrak{R}}}) , as the supremum of all integers M\in{\mathbb{N}} for which there is an increasing sequence 0<u_{1}<v_{1}\leq u_{2}<v_{2}\leq\cdots\leq u_{M}<v_{M} with u_{i},v_{i}\in\mathbb{L} such that |A_{u_{l}}^{{\mathfrak{R}}}f(x)-A_{v_{l}}^{{\mathfrak{R}}}f(x)|>\lambda holds for each l=1,\dots,M . We refer the readers to [ 11 ] for various properties involving \lambda- jumps. The following lemma states that the \ell^{2}- jump inequality implies the \ell^{2}- boundedness of the r- variation operator for r>2 . The following lemma is the discrete analogue of [ 11 , Lemma 2.1] . Since the proof of this lemma is exactly the same as that of [ 11 , Lemma 2.1] , we skip the details. Lemma 1.2 . [ 11 , Lemma 2.1] Let \mathcal{T} be a family of linear operators that maps \ell^{2}({\mathbb{Z}}^{n}) to the set of measurable functions on {\mathbb{Z}}^{n} . Suppose \displaystyle\sup_{\lambda>0}\|\lambda(\mathcal{J}_{\lambda}(\mathcal{T}f))^{1/2}\|_{\ell^{2}} \displaystyle\lesssim\|f\|_{\ell^{2}} holds for all f\in\ell^{2}({\mathbb{Z}}^{n}) . Then we have for r>2 , \displaystyle\|V_{r}(\mathcal{T}f)\|_{\ell^{2}} \displaystyle\lesssim\|f\|_{\ell^{2}}, for all f\in\ell^{2}({\mathbb{Z}}^{n}) . Therefore, Theorem 1.1 will be a consequence of the following result. Theorem 1.3 . Let \mathbb{L}\subset\Lambda be a lacunary sequence. Then, we have \displaystyle\sup_{\lambda>0}\|\lambda(\mathcal{J}_{\lambda}(\mathcal{A}_{\mathbb{L}}^{\mathfrak{R}}f))^{1/2}\|_{\ell^{2}} \displaystyle\lesssim\|f\|_{\ell^{2}}, for all f\in\ell^{2}({\mathbb{Z}}^{n}) . The proof of Theorem 1.3 is based on the ideas from [ 11 ] , where the jump

Read the complete text