Sampling in Graphs node sparsifiers Alexandr Andoni Microsoft Slides: 19 Download presentation Sampling in Graphs: node sparsifiers Alexandr Andoni (Microsoft Research) Graph compression Why smaller graphs? • use less storage space • faster algorithms • easier visualization Sparsification of edges • Preserve some structure: e. g. , cuts • Also: distances, effective resistances, etc Sparsification of nodes ? • Node sparsifiers • Results on cut sparsifiers Graph size Approximation Reference Comments [Moi 09, LM 10, CLLM 10, EGKRTCT 10, MM 10] [LM 10, CLLM 10, MM 10] [Chu 12, KW 12] [HKNR 98, KRTV 12] [KRTV 12, KR 13] bipartite* graphs [AGK’ 14] bipartite* graphs • Similar results for flow (node) sparsifier Small cut (node) sparsifiers [A-Gupta-Krauthgamer’ 14] • Main idea ? • Sampling edges doesn’t work here • Need to sample entire sub-structures of the graph Sampling in Bipartite Graphs • Sample non-terminals, together with edges • reweight edges accordingly Sampling in Bipartite Graphs • Sample non-terminals, together with edges • reweight edges accordingly • Uniform sampling doesn’t work Non-uniform sampling • Tool: Importance sampling • Importance sampling • Actual Sampling • Checking importance sampling • Flow (node) sparsifiers • Remarks • Graph compression via sampling • Alexandr andoniAlexandr andoniAlexandr veliký prezentaceNode and anti nodeNodenextReference node and non reference nodeRadial node angular nodeReference node and non reference nodeStruct node int data struct node* nextAlex andoniAlex andoniSpeed and velocityGraphs that enlighten and graphs that deceiveEnd behavior of polynomialsEncoding bugs in software testingStratified versus cluster samplingNatural sampling vs flat top samplingPerbedaan time sampling dan event samplingProbability sampling vs non probability samplingCluster sampling vs stratified random sampling