<p>We prove that Boolean matrices with bounded <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(\gamma _2\)</EquationSource> <EquationSource Format="MATHML"><math> <msub> <mi>γ</mi> <mn>2</mn> </msub> </math></EquationSource> </InlineEquation>-norm or bounded normalized trace norm must contain a linear-sized all-ones or all-zeros submatrix, verifying a conjecture of Hambardzumyan, Hatami, and Hatami. We also present further structural results about Boolean matrices of bounded <InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(\gamma _2\)</EquationSource> <EquationSource Format="MATHML"><math> <msub> <mi>γ</mi> <mn>2</mn> </msub> </math></EquationSource> </InlineEquation>-norm and discuss applications in communication complexity, operator theory, spectral graph theory, and extremal combinatorics. As a key application, we establish an inverse theorem for MaxCut. A celebrated result of Edwards states that every graph <i>G</i> with <i>m</i> edges has a cut of size at least <InlineEquation ID="IEq3"> <EquationSource Format="TEX">\(\frac{m}{2}+\frac{\sqrt{8m+1}-1}{8}\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mfrac> <mi>m</mi> <mn>2</mn> </mfrac> <mo>+</mo> <mfrac> <mrow> <msqrt> <mrow> <mn>8</mn> <mi>m</mi> <mo>+</mo> <mn>1</mn> </mrow> </msqrt> <mo>-</mo> <mn>1</mn> </mrow> <mn>8</mn> </mfrac> </mrow> </math></EquationSource> </InlineEquation>, with equality achieved by complete graphs with an odd number of vertices. To contrast this, we prove that if the MaxCut of <i>G</i> is at most <InlineEquation ID="IEq4"> <EquationSource Format="TEX">\(\frac{m}{2}+O(\sqrt{m})\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mfrac> <mi>m</mi> <mn>2</mn> </mfrac> <mo>+</mo> <mi>O</mi> <mrow> <mo stretchy="false">(</mo> <msqrt> <mi>m</mi> </msqrt> <mo stretchy="false">)</mo> </mrow> </mrow> </math></EquationSource> </InlineEquation>, then <i>G</i> must contain a clique of size <InlineEquation ID="IEq5"> <EquationSource Format="TEX">\(\Omega (\sqrt{m})\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi mathvariant="normal">Ω</mi> <mo stretchy="false">(</mo> <msqrt> <mi>m</mi> </msqrt> <mo stretchy="false">)</mo> </mrow> </math></EquationSource> </InlineEquation>.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Factorization norms and an inverse theorem for MaxCut

  • Igor Balla,
  • Lianna Hambardzumyan,
  • István Tomon

摘要

We prove that Boolean matrices with bounded \(\gamma _2\) γ 2 -norm or bounded normalized trace norm must contain a linear-sized all-ones or all-zeros submatrix, verifying a conjecture of Hambardzumyan, Hatami, and Hatami. We also present further structural results about Boolean matrices of bounded \(\gamma _2\) γ 2 -norm and discuss applications in communication complexity, operator theory, spectral graph theory, and extremal combinatorics. As a key application, we establish an inverse theorem for MaxCut. A celebrated result of Edwards states that every graph G with m edges has a cut of size at least \(\frac{m}{2}+\frac{\sqrt{8m+1}-1}{8}\) m 2 + 8 m + 1 - 1 8 , with equality achieved by complete graphs with an odd number of vertices. To contrast this, we prove that if the MaxCut of G is at most \(\frac{m}{2}+O(\sqrt{m})\) m 2 + O ( m ) , then G must contain a clique of size \(\Omega (\sqrt{m})\) Ω ( m ) .