0.25
0.5
0.75
1.25
1.5
1.75
2
PhD Thesis Defense: Dynamics of large networks
Published on Oct 22, 2008137746 Views
A basic premise behind the study of large networks is that interaction leads to complex collective behavior. In our work we found very interesting and counterintuitive patterns for time evolving net
Related categories
Chapter list
Dynamics of large networks00:00
Web: Rich data00:59
Rich data: Networks02:08
Networks. What do we know?02:50
This thesis: Network dynamics03:46
This thesis. The structure (1)04:38
This thesis. The structure (2)05:17
Background: Network models05:20
Q1) Network evolution (1)06:49
Q1) Network evolution (2)07:52
Q2) Modeling edge attachment08:43
Setting: Edge-by-edge evolution10:06
Edge attachment degree bias11:25
But, edges also attach locally12:44
How to best close a triangle?14:20
Q3) Generating realistic graphs16:13
Q3) The model: Kronecker graphs17:32
Q5) Kronecker graphs: Estimation19:03
Estimation: Epinions (N=76k, E=510k)21:04
Thesis: The structure (1)21:58
Thesis: The structure (2)22:13
Part 2. Diffusion and Cascades22:14
Settting 1: Viral marketing23:24
Setting 2: Blogosphere24:19
Q4) What do cascades look like?25:17
Q5) Human adoption curves27:05
Q5) Adoption curve: Validation28:09
Q6) Cascade & outbreak detection28:52
Q6) The problem: Detecting cascades29:47
Two parts to the problem30:10
Optimization problem31:05
Solution: CELF Algorithm31:38
Problem structure: Submodularity32:11
Blogs: Information epidemics (1)33:17
Blogs: Information epidemics (2)34:18
CELF: Scalability34:47
Same problem: Water Network35:15
Water network: Results35:55
Thesis: The structure (3)36:31
Thesis: The structure (4)36:53
3 case studies on large data36:54
Q7) Planetary look on small-world37:34
Q8) Network community structure40:06
Example: Small network41:09
Example: Large network42:03
Q8) Suggested network structure (1)42:41
Q8) Suggested network structure (2)43:39
Q9) Web projections44:29
Q9) Web projections: Results45:36
Thesis: The structure (5)46:49
Future directions: Evolution47:05
Future directions: Diffusion48:31
What's next?49:33