<?xml version="1.0" encoding="UTF-8"?>
<OAI-PMH xmlns="http://www.openarchives.org/OAI/2.0/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/ http://www.openarchives.org/OAI/2.0/OAI-PMH.xsd">
  <responseDate>2026-07-24T00:01:39Z</responseDate>
  <request identifier="7580" metadataPrefix="oai_dc" verb="GetRecord">https://drops.dagstuhl.de/oai</request>
  <GetRecord>
    <record>
      <header>
        <identifier>oai:drops-oai.dagstuhl.de:7580</identifier>
        <datestamp>2024-03-06T10:40:40Z</datestamp>
        <setSpec>ddc:004</setSpec>
        <setSpec>open_access</setSpec>
      </header>
      <metadata>
        <oai_dc:dc xmlns:oai_dc="http://www.openarchives.org/OAI/2.0/oai_dc/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/oai_dc/ http://www.openarchives.org/OAI/2.0/oai_dc.xsd">
          <dc:title>Sum-of-Squares Certificates for Maxima of Random Tensors on the Sphere</dc:title>
          <dc:creator>Bhattiprolu, Vijay</dc:creator>
          <dc:creator>Guruswami, Venkatesan</dc:creator>
          <dc:creator>Lee, Euiwoong</dc:creator>
          <dc:subject>Sum-of-Squares</dc:subject>
          <dc:subject>Optimization over Sphere</dc:subject>
          <dc:subject>Random Polynomials</dc:subject>
          <dc:description>For an n-variate order-d tensor A, define A_{max} := sup_{||x||_2 = 1} &lt;A,x^(otimes d)&gt;, to be the maximum value taken by the tensor on the unit sphere. It is known that for a random tensor with i.i.d. +1/-1 entries, A_{max} &lt;= sqrt(n.d.log(d)) w.h.p. We study the problem of efficiently certifying upper bounds on A_{max} via the natural relaxation from the Sum of Squares (SoS) hierarchy. Our results include:&#13;
&#13;
* When A is a random order-q tensor, we prove that q levels of SoS certifies an upper bound B on A_{max} that satisfies B &lt;= A_{max} * (n/q^(1-o(1)))^(q/4-1/2) w.h.p. Our upper bound improves a result of Montanari and Richard (NIPS 2014) when q is large.&#13;
&#13;
* We show the above bound is the best possible up to lower order terms, namely the optimum of the level-q SoS relaxation is at least &#13;
A_{max} * (n/q^(1+o(1)))^(q/4-1/2). &#13;
&#13;
* When A is a random order-d tensor, we prove that q levels of SoS certifies an upper bound B on A_{max} that satisfies B &lt;= A_{max} * (n*polylog/q)^(d/4 - 1/2) w.h.p. For growing q, this improves upon the bound certified by constant levels of SoS. This answers in part, a question posed by Hopkins, Shi, and Steurer (COLT 2015), who tightly characterized constant levels of SoS.</dc:description>
          <dc:publisher>Schloss Dagstuhl – Leibniz-Zentrum für Informatik</dc:publisher>
          <dc:contributor>Vijay Bhattiprolu and Venkatesan Guruswami and Euiwoong Lee</dc:contributor>
          <dc:date>2017</dc:date>
          <dc:relation>Is Part Of LIPIcs, Volume 81, Approximation, Randomization, and Combinatorial Optimization. Algorithms and Techniques (APPROX/RANDOM 2017)</dc:relation>
          <dc:type>InProceedings</dc:type>
          <dc:type>Text</dc:type>
          <dc:type>doc-type:ResearchArticle</dc:type>
          <dc:type>publishedVersion</dc:type>
          <dc:format>application/pdf</dc:format>
          <dc:identifier>doi:10.4230/LIPIcs.APPROX-RANDOM.2017.31</dc:identifier>
          <dc:identifier>urn:nbn:de:0030-drops-75808</dc:identifier>
          <dc:identifier>https://drops.dagstuhl.de/entities/document/10.4230/LIPIcs.APPROX-RANDOM.2017.31</dc:identifier>
          <dc:language>eng</dc:language>
          <dc:rights>https://creativecommons.org/licenses/by/3.0/legalcode</dc:rights>
        </oai_dc:dc>
      </metadata>
    </record>
  </GetRecord>
</OAI-PMH>
