Sign inSign up
Edge Label Inference in Generalized Stochastic Block Models: from Spectral Theory to Impossibility Results thumbnail
Pause
Mute
Subtitles
Playback speed
0.25
0.5
0.75
1
1.25
1.5
1.75
2
Full screen

Edge Label Inference in Generalized Stochastic Block Models: from Spectral Theory to Impossibility Results

Published on Jul 15, 20142124 Views

Laurent Massouile

The classical setting of community detection consists of networks exhibiting a clustered structure. To more accurately model real systems we consider a class of networks (i) whose edges may carry labe

COLT 2014 - Barcelona

Related categories

Computational Learning TheoryOn-line LearningStatistical LearningSupervised LearningUnsupervised Learning

Chapter list

Edge Label inference in generalized stochastic block models 00:00
The Stochastic Block Model [Holland-Laskey-Leinhardt’83]00:06
SBM with labels01:22
SBM with general types [Bollobas-Janson-Riordan’07,…]01:50
Pre-processing: random label projections02:45
Spectral properties for logarithmic S [tools: Kolchinski’98, Feige-Ofek’05]03:25
Label inference for logarithmic S04:29
Consistent estimation of label distributions - 104:52
Consistent estimation of label distributions - 204:59
Consistent estimation of label distributions - 305:04
Thank you05:24