2018
DOI: 10.1007/978-3-030-04693-4_17
|View full text |Cite
|
Sign up to set email alerts
|

Cut Sparsifiers for Balanced Digraphs

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
3
2

Citation Types

2
10
0

Year Published

2019
2019
2024
2024

Publication Types

Select...
4
1

Relationship

0
5

Authors

Journals

citations
Cited by 6 publications
(12 citation statements)
references
References 15 publications
2
10
0
Order By: Relevance
“…Our lower bound construction also refutes a conjecture of Ikeda and Tanigawa [20] that replacing β/λ e with the tighter 1/γ e in the sampling probabilities p e , where γ e is the directed edge connectivity of e, produces a better sparsifier. We give an example to show that sampling with probabilities proportional to 1/γ e may not even produce a sparsifier, i.e., does not preserve the values of all cuts!…”
Section: Our Resultssupporting
confidence: 83%
See 4 more Smart Citations
“…Our lower bound construction also refutes a conjecture of Ikeda and Tanigawa [20] that replacing β/λ e with the tighter 1/γ e in the sampling probabilities p e , where γ e is the directed edge connectivity of e, produces a better sparsifier. We give an example to show that sampling with probabilities proportional to 1/γ e may not even produce a sparsifier, i.e., does not preserve the values of all cuts!…”
Section: Our Resultssupporting
confidence: 83%
“…"For-All" Sparsification. Recently, Ikeda and Tanigawa [20] showed that there is a sparsifier for β-balanced directed graphs that uses O(βn/ 2 ) edges. 1 Indeed, this result can be recovered by using the following simple process: (a) remove all edge directions, (b) sample the resulting undirected graph using a standard (undirected) sparsification algorithm that samples every edge e at probability p e that is inversely proportional to the (undirected) edge connectivity λ e (see [17]), but after boosting the sampling probabilities by a factor of β, and (c) use a simple modification of the analysis from undirected sparsification to argue that the directed cuts are preserved because of the higher sampling probabilities.…”
Section: Our Resultsmentioning
confidence: 99%
See 3 more Smart Citations