FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2563182657> ?p ?o ?g. }

• W2563182657 abstract "A counterexample to a strong variant of the polynomial Freiman-Ruzsa conjecture, Discrete Analysis 2017:8, 6 pp. Given a finite set $A$ of integers, define its _sumset_ $A+A$ to be the set ${x+y:x,yin A}$. A central question in additive combinatorics is the following: what can we say about a set $A$ if its sumset is small? A remarkable theorem of Freiman, later given a different and very influential proof by Imre Ruzsa, gives a complete answer to this question, at least qualitatively speaking, when small is interpreted as of size at most $C|A|$ for an absolute constant $C$. An easy observation is that if $A$ is an arithmetic progression, then $|A+A|=2|A|-1$. Another easy observation is that if $|A+A|leq C|A|$ and $Bsubset A$ is a subset with $|B|geq c|A|$, then $|B+B|leq c^{-1}C|B|$. A slightly less easy observation is that higher-dimensional progressions have small sumsets: for example, a set $A$ of the form ${a_0+d_1x+d_2y: 0 leq x < m_1, 0 leq y < m_2 }$, which is a 2-dimensional arithmetic progression, has a sumset of size at most $4|A|$, roughly speaking because it expands by a factor 2 in each direction. Freiman's theorem is the much less obvious statement that these observations give all examples: every set with a small sumset is a large subset of a low-dimensional arithmetic progression. Freiman's original argument gave very weak bounds. Although these were substantially improved by Ruzsa, and improved further by Chang, the theorem nevertheless had the following drawback. If you start with a set $A$ with a sumset of size at most $C|A|$ and apply Freiman's theorem, then you obtain the information that $A$ is contained in an arithmetic progression of size at most $K|A|$ and dimension at most $d$, for constants $K$ and $d$ that depend on $C$. But if you use just this information and go back in the other direction, then all you can deduce is that such a set has sumset of size at most $2^dK|A|$, and even with Chang's bounds for $d$ and $K$, $2^dK$ is far bigger than $C$. Thus, there is a sense in which Freiman's theorem does not capture the full structure of a set $A$ with small sumset. This situation was radically improved in 2010 by Sanders, who obtained a quasipolynomial version of the theorem: that is, if you apply the theorem and its trivial converse, you recover a constant that is at most $exp((log C)^t)$ for an absolute constant $t$. However, the holy grail is to obtain a characterization of sets with small sumset (or, as they are often called, sets with small doubling), that worsens the constant by at most a polynomial. The polynomial Freiman-Ruzsa conjecture would, if true, give a characterization of this kind. To state it, we need the notion of a _symmetric convex lattice set_: that is, the intersection of a lattice with a symmetric convex body. If a convex lattice set has dimension $d$, it can be shown that its doubling constant is at most $5^d$. One can project a convex lattice set into a lower dimension and preserve this property, which gives a class of sets of small doubling that is more general than multidimensional arithmetic progressions: a multidimensional arithmetic progression is a projected convex lattice set for which the lattice is $mathbb Z^d$ and the convex body intersecting it is an axis-aligned cuboid. If $B$ is a projected convex lattice set of bounded dimension, and $X$ is a set of bounded size, then $B+X$ still has small doubling, and the polynomial Freiman-Ruzsa conjecture is roughly that every set $A$ with $|A+A|leq C|A|$ is contained in such a set, with the sizes of $B$ and $X$ and the dimension of the lattice being such that from this information alone one can recover a doubling constant of $P(C)$ for some fixed polynomial $P$. Above, we observed that projections of convex lattice sets give us a more general class of sets of small doubling than multidimensional arithmetic progressions. But do we need this extra generality, given that we can also pass to large subsets? In a qualitative sense, we do not, by Freiman's theorem (or by a direct argument using Minkowski's second theorem). However, if we want a characterization that gives only a polynomial loss, then it is not obvious whether the extra generality is needed. That is the question addressed by this paper. (The paper deals with subsets of $mathbb R^n$ rather than sets of integers but this makes no difference.) The authors prove that if one strengthens the polynomial Freiman-Ruzsa conjecture by insisting that the projected lattice convex bodies are multidimensional arithmetic progressions, then the resulting statement is false. This answers a question of Ben Green. In some ways, the result is not surprising: Green expected the strengthened statement to be false, and if it is to be false, then the natural place to look for counterexamples is convex lattice sets that are as dissimilar as possible to aligned cuboids. So the authors' counterexample -- the intersection of a random lattice with a Euclidean ball -- is a natural one. However, it is not clear how to prove that examples of this kind really are counterexamples. The main achievement of this paper is to observe that a recent result (a reverse Minkowski theorem due to the second author and Noah Stephens-Davidowitz) can be used to give a short proof, thereby confirming that the statement of the polynomial Freiman-Ruzsa conjecture is the correct one." @default.
• W2563182657 created "2017-01-06" @default.
• W2563182657 creator A5044999305 @default.
• W2563182657 creator A5049700941 @default.
• W2563182657 date "2018-01-01" @default.
• W2563182657 modified "2023-10-16" @default.
• W2563182657 title "A counterexample to a strong variant of the Polynomial Freiman-Ruzsa conjecture in Euclidean space" @default.
• W2563182657 doi "https://doi.org/10.19086/da.1640" @default.
• W2563182657 hasPublicationYear "2018" @default.
• W2563182657 type Work @default.
• W2563182657 sameAs 2563182657 @default.
• W2563182657 citedByCount "19" @default.
• W2563182657 countsByYear W25631826572016 @default.
• W2563182657 countsByYear W25631826572017 @default.
• W2563182657 countsByYear W25631826572019 @default.
• W2563182657 countsByYear W25631826572020 @default.
• W2563182657 countsByYear W25631826572021 @default.
• W2563182657 countsByYear W25631826572023 @default.
• W2563182657 crossrefType "journal-article" @default.
• W2563182657 hasAuthorship W2563182657A5044999305 @default.
• W2563182657 hasAuthorship W2563182657A5049700941 @default.
• W2563182657 hasBestOaLocation W25631826571 @default.
• W2563182657 hasConcept C111919701 @default.
• W2563182657 hasConcept C114614502 @default.
• W2563182657 hasConcept C118615104 @default.
• W2563182657 hasConcept C129782007 @default.
• W2563182657 hasConcept C130498168 @default.
• W2563182657 hasConcept C134306372 @default.
• W2563182657 hasConcept C162838799 @default.
• W2563182657 hasConcept C176370821 @default.
• W2563182657 hasConcept C177264268 @default.
• W2563182657 hasConcept C185592680 @default.
• W2563182657 hasConcept C186450821 @default.
• W2563182657 hasConcept C199360897 @default.
• W2563182657 hasConcept C2524010 @default.
• W2563182657 hasConcept C2778572836 @default.
• W2563182657 hasConcept C2780990831 @default.
• W2563182657 hasConcept C33923547 @default.
• W2563182657 hasConcept C41008148 @default.
• W2563182657 hasConcept C55493867 @default.
• W2563182657 hasConcept C90119067 @default.
• W2563182657 hasConcept C94375191 @default.
• W2563182657 hasConcept C98184364 @default.
• W2563182657 hasConceptScore W2563182657C111919701 @default.
• W2563182657 hasConceptScore W2563182657C114614502 @default.
• W2563182657 hasConceptScore W2563182657C118615104 @default.
• W2563182657 hasConceptScore W2563182657C129782007 @default.
• W2563182657 hasConceptScore W2563182657C130498168 @default.
• W2563182657 hasConceptScore W2563182657C134306372 @default.
• W2563182657 hasConceptScore W2563182657C162838799 @default.
• W2563182657 hasConceptScore W2563182657C176370821 @default.
• W2563182657 hasConceptScore W2563182657C177264268 @default.
• W2563182657 hasConceptScore W2563182657C185592680 @default.
• W2563182657 hasConceptScore W2563182657C186450821 @default.
• W2563182657 hasConceptScore W2563182657C199360897 @default.
• W2563182657 hasConceptScore W2563182657C2524010 @default.
• W2563182657 hasConceptScore W2563182657C2778572836 @default.
• W2563182657 hasConceptScore W2563182657C2780990831 @default.
• W2563182657 hasConceptScore W2563182657C33923547 @default.
• W2563182657 hasConceptScore W2563182657C41008148 @default.
• W2563182657 hasConceptScore W2563182657C55493867 @default.
• W2563182657 hasConceptScore W2563182657C90119067 @default.
• W2563182657 hasConceptScore W2563182657C94375191 @default.
• W2563182657 hasConceptScore W2563182657C98184364 @default.
• W2563182657 hasLocation W25631826571 @default.
• W2563182657 hasLocation W25631826572 @default.
• W2563182657 hasLocation W25631826573 @default.
• W2563182657 hasOpenAccess W2563182657 @default.
• W2563182657 hasPrimaryLocation W25631826571 @default.
• W2563182657 hasRelatedWork W1969822793 @default.
• W2563182657 hasRelatedWork W2123831997 @default.
• W2563182657 hasRelatedWork W2563182657 @default.
• W2563182657 hasRelatedWork W2951096767 @default.
• W2563182657 hasRelatedWork W3153743757 @default.
• W2563182657 hasRelatedWork W3194430607 @default.
• W2563182657 hasRelatedWork W4200431641 @default.
• W2563182657 hasRelatedWork W4287024420 @default.
• W2563182657 hasRelatedWork W4293869835 @default.
• W2563182657 hasRelatedWork W4297917402 @default.
• W2563182657 isParatext "false" @default.
• W2563182657 isRetracted "false" @default.
• W2563182657 magId "2563182657" @default.
• W2563182657 workType "article" @default.