Hitting sets and reconstruction for dense orbits in VPe and ΣΠΣ circuits

D. Medini, Amir Shpilka
{"title":"Hitting sets and reconstruction for dense orbits in VPe and ΣΠΣ circuits","authors":"D. Medini, Amir Shpilka","doi":"10.4230/LIPIcs.CCC.2021.19","DOIUrl":null,"url":null,"abstract":"In this paper we study polynomials in VPe (polynomial-sized formulas) and in ΣΠΣ (polynomial-size depth-3 circuits) whose orbits, under the action of the affine group GLaffn (F) (the action of (A, b) ∈ GLaffn (F) on a polynomial f ∈ F[x] is defined as (A, b) ο f = f (AT x + b)), are dense in their ambient class. We construct hitting sets and interpolating sets for these orbits as well as give reconstruction algorithms. Specifically, we obtain the following results: 1. For [EQUATION], where the ℓis are linearly independent linear functions, we construct a polynomial-sized interpolating set, and give a polynomial-time reconstruction algorithm. By a result of Bringmann, Ikenmeyer and Zuiddam, the set of all such polynomials is dense in VPe [14], thus our construction gives the first polynomial-size interpolating set for a dense subclass of VPe. 2. For polynomials of the form ANFΔ (ℓ1(x),..., ℓ4Δ(x)) where ANFΔ(x) is the canonical read-once formula in alternating normal form, of depth 2Δ, and the ℓis are linearly independent linear functions, we provide a quasipolynomial-size interpolating set. We also observe that the reconstruction algorithm of [35] works for all polynomials in this class. This class is also dense in VPe. 3. Similarly, we give a quasipolynomial-sized hitting set for read-once formulas (not necessarily in alternating normal form) composed with a set of linearly independent linear functions. This gives another dense class in VPe. 4. We give a quasipolynomial-sized hitting set for polynomials of the form f(ℓ1(x),..., ℓm(x)), where f is an m-variate s-sparse polynomial. and the ℓis are linearly independent linear functions in n ≥ m variables. This class is dense in ΣΠΣ. 5. For polynomials of the form Σsi=1 Πdj=1 ℓi,j (x), where the ℓi,js are linearly independent linear functions, we construct a polynomial-sized interpolating set. We also observe that the reconstruction algorithm of [45] works for every polynomial in the class. This class is dense in ΣΠΣ. As VP = VNC2, our results for VPe translate immediately to VP with a quasipolynomial blow up in parameters. If any of our hitting or interpolating sets could be made robust then this would immediately yield a hitting set for the superclass in which the relevant class is dense, and as a consequence also a lower bound for the superclass. Unfortunately, we also prove that the kind of constructions that we have found (which are defined in terms of k-independent polynomial maps) do not necessarily yield robust hitting sets.","PeriodicalId":336911,"journal":{"name":"Proceedings of the 36th Computational Complexity Conference","volume":"27 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2021-07-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"10","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings of the 36th Computational Complexity Conference","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.4230/LIPIcs.CCC.2021.19","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 10

Abstract

In this paper we study polynomials in VPe (polynomial-sized formulas) and in ΣΠΣ (polynomial-size depth-3 circuits) whose orbits, under the action of the affine group GLaffn (F) (the action of (A, b) ∈ GLaffn (F) on a polynomial f ∈ F[x] is defined as (A, b) ο f = f (AT x + b)), are dense in their ambient class. We construct hitting sets and interpolating sets for these orbits as well as give reconstruction algorithms. Specifically, we obtain the following results: 1. For [EQUATION], where the ℓis are linearly independent linear functions, we construct a polynomial-sized interpolating set, and give a polynomial-time reconstruction algorithm. By a result of Bringmann, Ikenmeyer and Zuiddam, the set of all such polynomials is dense in VPe [14], thus our construction gives the first polynomial-size interpolating set for a dense subclass of VPe. 2. For polynomials of the form ANFΔ (ℓ1(x),..., ℓ4Δ(x)) where ANFΔ(x) is the canonical read-once formula in alternating normal form, of depth 2Δ, and the ℓis are linearly independent linear functions, we provide a quasipolynomial-size interpolating set. We also observe that the reconstruction algorithm of [35] works for all polynomials in this class. This class is also dense in VPe. 3. Similarly, we give a quasipolynomial-sized hitting set for read-once formulas (not necessarily in alternating normal form) composed with a set of linearly independent linear functions. This gives another dense class in VPe. 4. We give a quasipolynomial-sized hitting set for polynomials of the form f(ℓ1(x),..., ℓm(x)), where f is an m-variate s-sparse polynomial. and the ℓis are linearly independent linear functions in n ≥ m variables. This class is dense in ΣΠΣ. 5. For polynomials of the form Σsi=1 Πdj=1 ℓi,j (x), where the ℓi,js are linearly independent linear functions, we construct a polynomial-sized interpolating set. We also observe that the reconstruction algorithm of [45] works for every polynomial in the class. This class is dense in ΣΠΣ. As VP = VNC2, our results for VPe translate immediately to VP with a quasipolynomial blow up in parameters. If any of our hitting or interpolating sets could be made robust then this would immediately yield a hitting set for the superclass in which the relevant class is dense, and as a consequence also a lower bound for the superclass. Unfortunately, we also prove that the kind of constructions that we have found (which are defined in terms of k-independent polynomial maps) do not necessarily yield robust hitting sets.
VPe和ΣΠΣ电路中密集轨道的撞击集和重建
本文研究了VPe(多项式大小的公式)和ΣΠΣ(多项式大小的深度-3电路)中的多项式,它们的轨道在仿射群GLaffn (F)的作用下((A, b)∈GLaffn (F)对多项式F∈F[x]的作用定义为(A, b) ο F = F (AT x + b))在其环境类中是密集的。我们构造了这些轨道的命中集和插值集,并给出了重建算法。具体而言,我们得到以下结果:1。对于[EQUATION],其中的l是线性无关的线性函数,我们构造了一个多项式大小的插值集,并给出了一个多项式时间重构算法。根据Bringmann, Ikenmeyer和Zuiddam的结果,所有这些多项式的集合在VPe[14]中是密集的,因此我们的构造给出了VPe的密集子类的第一个多项式大小的插值集。2. 对于形式为ANFΔ (p1 (x),…,其中ANFΔ(x)是深度为2Δ的交替正规形式的标准读一次公式,并且是线性无关的线性函数,我们提供了一个拟多项式大小的插值集。我们还观察到[35]的重构算法适用于该类中的所有多项式。这个类在VPe中也很密集。3.类似地,我们给出了一个拟多项式大小的命中集,用于由一组线性无关的线性函数组成的一次读取公式(不一定是交替正规形式)。这在VPe中提供了另一个密集的类。4. 我们给出了一个拟多项式大小的撞击集,其形式为f(1(x),…), m(x)),其中f是一个m变量s稀疏多项式。并且,在n≥m个变量中,是线性无关的线性函数。这个类是密集的ΣΠΣ。5. 对于形式为Σsi=1 Πdj=1的多项式,其中的i,js是线性无关的线性函数,我们构造了一个多项式大小的插值集。我们还观察到[45]的重构算法适用于该类中的每个多项式。这个类是密集的ΣΠΣ。当VP = VNC2时,我们对VPe的结果立即转化为具有准多项式参数膨胀的VP。如果我们的任何命中集或插值集可以变得鲁棒,那么这将立即产生超类的命中集,其中相关类是密集的,因此也是超类的下界。不幸的是,我们也证明了我们所发现的那种构造(用k无关的多项式映射来定义)并不一定产生鲁棒的命中集。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
自引率
0.00%
发文量
0
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:604180095
Book学术官方微信