paper

Algebraic Methods in the Congested Clique

arXiv:1503.04963 · doi:10.1007/s00446-016-0270-2

Abstract

In this work, we use algebraic methods for studying distance computation and subgraph detection tasks in the congested clique model. Specifically, we adapt parallel matrix multiplication implementations to the congested clique, obtaining an round matrix multiplication algorithm, where is the exponent of matrix multiplication. In conjunction with known techniques from centralised algorithmics, this gives significant improvements over previous best upper bounds in the congested clique model. The highlight results include: -- triangle and 4-cycle counting in rounds, improving upon the triangle detection algorithm of Dolev et al. [DISC 2012], -- a -approximation of all-pairs shortest paths in rounds, improving upon the -round -approximation algorithm of Nanongkai [STOC 2014], and -- computing the girth in rounds, which is the first non-trivial solution in this model. In addition, we present a novel constant-round combinatorial algorithm for detecting 4-cycles.

This is work is a merger of arxiv:1412.2109 and arxiv:1412.2667

References in corpus (4)

Cited by in corpus (21)