Hadamard Tensors and Lower Bounds on Multiparty Communication Complexity

Authors Jeff Ford, Anna Gál



PDF
Thumbnail PDF

File

DagSemProc.06111.9.pdf
  • Filesize: 215 kB
  • 20 pages

Document Identifiers

Author Details

Jeff Ford
Anna Gál

Cite AsGet BibTex

Jeff Ford and Anna Gál. Hadamard Tensors and Lower Bounds on Multiparty Communication Complexity. In Complexity of Boolean Functions. Dagstuhl Seminar Proceedings, Volume 6111, pp. 1-20, Schloss Dagstuhl – Leibniz-Zentrum für Informatik (2006)
https://doi.org/10.4230/DagSemProc.06111.9

Abstract

We develop a new method for estimating the discrepancy of tensors associated with multiparty communication problems in the ``Number on the Forehead'' model of Chandra, Furst and Lipton. We define an analogue of the Hadamard property of matrices for tensors in multiple dimensions and show that any $k$-party communication problem represented by a Hadamard tensor must have $Omega(n/2^k)$ multiparty communication complexity. We also exhibit constructions of Hadamard tensors, giving $Omega(n/2^k)$ lower bounds on multiparty communication complexity for a new class of explicitly defined Boolean functions.
Keywords
  • Multiparty communication complexity
  • lower bounds

Metrics

  • Access Statistics
  • Total Accesses (updated on a weekly basis)
    0
    PDF Downloads
Questions / Remarks / Feedback
X

Feedback for Dagstuhl Publishing


Thanks for your feedback!

Feedback submitted

Could not send message

Please try again later or send an E-mail