Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Graph reduction with spectral and cut guarantees

About

Can one reduce the size of a graph without significantly altering its basic properties? The graph reduction problem is hereby approached from the perspective of restricted spectral approximation, a modification of the spectral similarity measure used for graph sparsification. This choice is motivated by the observation that restricted approximation carries strong spectral and cut guarantees, and that it implies approximation results for unsupervised learning problems relying on spectral embeddings. The paper then focuses on coarsening---the most common type of graph reduction. Sufficient conditions are derived for a small graph to approximate a larger one in the sense of restricted similarity. These findings give rise to nearly-linear algorithms that, compared to both standard and advanced graph reduction methods, find coarse graphs of improved quality, often by a large margin, without sacrificing speed.

Andreas Loukas• 2018

Related benchmarks

TaskDatasetResultRank
Graph CoarseningCiteseer
Running Time (s)9.58
17
Graph CoarseningComputers
Running Time (s)39.33
11
Graph CoarseningPubmed
Running Time (s)125
11
Graph CoarseningCo-Cs
Running Time (s)218.1
11
Graph CoarseningCo-phy
Running Time (s)466
11
Graph CoarseningPhoto
Running Time (s)27.15
11
Graph CoarseningCora
Running Time (s)9.04
11
Graph CoarseningFlickr
Running Time (s)314.6
10
Graph CoarseningREDDIT
Running Time (s)943.9
8
Showing 9 of 9 rows

Other info

Follow for update