en
0.25
0.5
0.75
1.25
1.5
1.75
2
Fast Influence-based Coarsening for Large Networks
Published on Oct 07, 20141609 Views
Given a social network, can we quickly 'zoom-out' of the graph? Is there a smaller equivalent representation of the graph that preserves its propagation characteristics? Can we group nodes together ba
Related categories
Chapter list
Fast Influence-based Coarsening for Large Networks00:00
Networks are getting huge!00:12
Need for fast analysis00:30
How to handle large-scale networks00:52
Bird’s eye view of a network - 101:14
Bird’s eye view of a network - 201:29
Outline - 102:04
Challenges02:06
C1: Information Diffusion02:42
C1: Model information diffusion03:07
C1: Diffusive characteristics03:56
C1: maintain diffusive characteristics05:41
C2: How to merge nodes - 106:03
C2: How to merge nodes - 208:01
C3: which nodes to merge08:18
Outline - 208:41
Problem Definition08:41
Naive Greedy Heuristic09:16
Outline - 309:52
CoarseNet: idea09:54
CoarseNet: details10:10
CoarseNet10:31
CoarseNet: Complete algorithm10:47
CoarseNet: running time11:33
Outline - 411:38
How do we perform?11:44
Scalability w.r.t Graph Size (n)12:51
Scalability w.r.t Reduction Factor (α) - 112:53
Outline - 513:03
Application 1: Influence Maximization - 113:22
Application 1: Influence Maximization - 214:03
Quality of CSPIN15:40
Quality of CSPIN w.r.t α16:35
Scalability w.r.t number of seeds16:40
Application 2: Diffusion Characterization17:10
Diffusion observation17:59
Outline - 619:09
Conclusion19:10
Any Questions?19:34