![Page 1: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/1.jpg)
CMU SCS
Large Graph Mining - Patterns, Explanations and
Cascade Analysis Christos Faloutsos
CMU
![Page 2: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/2.jpg)
CMU SCS
Thank you!
• Michalis Vazirgiannis
• Mauro Sozio
DSBDE, July 2013 (c) 2013, C. Faloutsos 2
![Page 3: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/3.jpg)
CMU SCS
(c) 2013, C. Faloutsos 3
Roadmap
• Introduction – Motivation – Why study (big) graphs?
• Part#1: Patterns in graphs • Part#2: Cascade analysis • Conclusions • [Extra: ebay fraud; tensors; spikes]
DSBDE, July 2013
![Page 4: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/4.jpg)
CMU SCS
(c) 2013, C. Faloutsos 4
Graphs - why should we care?
Internet Map [lumeta.com]
Food Web [Martinez ’91]
>$10B revenue
>0.5B users
DSBDE, July 2013
![Page 5: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/5.jpg)
CMU SCS
(c) 2013, C. Faloutsos 5
Graphs - why should we care? • web-log (‘blog’) news propagation • computer network security: email/IP traffic and
anomaly detection • Recommendation systems • ....
• Many-to-many db relationship -> graph
DSBDE, July 2013
![Page 6: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/6.jpg)
CMU SCS
(c) 2013, C. Faloutsos 6
Roadmap
• Introduction – Motivation • Part#1: Patterns in graphs
– Static graphs – Time-evolving graphs – Why so many power-laws?
• Part#2: Cascade analysis • Conclusions
DSBDE, July 2013
![Page 7: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/7.jpg)
CMU SCS
DSBDE, July 2013 (c) 2013, C. Faloutsos 7
Part 1: Patterns & Laws
![Page 8: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/8.jpg)
CMU SCS
(c) 2013, C. Faloutsos 8
Laws and patterns • Q1: Are real graphs random?
DSBDE, July 2013
![Page 9: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/9.jpg)
CMU SCS
(c) 2013, C. Faloutsos 9
Laws and patterns • Q1: Are real graphs random? • A1: NO!!
– Diameter – in- and out- degree distributions – other (surprising) patterns
• Q2: why ‘no good cuts’? • A2: <self-similarity – stay tuned>
• So, let’s look at the data DSBDE, July 2013
![Page 10: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/10.jpg)
CMU SCS
(c) 2013, C. Faloutsos 10
Solution# S.1 • Power law in the degree distribution
[SIGCOMM99]
log(rank)
log(degree)
internet domains
att.com
ibm.com
DSBDE, July 2013
![Page 11: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/11.jpg)
CMU SCS
(c) 2013, C. Faloutsos 11
Solution# S.1 • Power law in the degree distribution
[SIGCOMM99]
log(rank)
log(degree)
-0.82
internet domains
att.com
ibm.com
DSBDE, July 2013
![Page 12: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/12.jpg)
CMU SCS
(c) 2013, C. Faloutsos 12
Solution# S.1 • Q: So what?
log(rank)
log(degree)
-0.82
internet domains
att.com
ibm.com
DSBDE, July 2013
![Page 13: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/13.jpg)
CMU SCS
(c) 2013, C. Faloutsos 13
Solution# S.1 • Q: So what? • A1: # of two-step-away pairs:
log(rank)
log(degree)
-0.82
internet domains
att.com
ibm.com
DSBDE, July 2013
= friends of friends (F.O.F.)
![Page 14: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/14.jpg)
CMU SCS
(c) 2013, C. Faloutsos 14
Solution# S.1 • Q: So what? • A1: # of two-step-away pairs: O(d_max ^2) ~ 10M^2
log(rank)
log(degree)
-0.82
internet domains
att.com
ibm.com
DSBDE, July 2013
~0.8PB -> a data center(!)
DCO @ CMU
Gaussian trap
= friends of friends (F.O.F.)
![Page 15: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/15.jpg)
CMU SCS
(c) 2013, C. Faloutsos 15
Solution# S.1 • Q: So what? • A1: # of two-step-away pairs: O(d_max ^2) ~ 10M^2
log(rank)
log(degree)
-0.82
internet domains
att.com
ibm.com
DSBDE, July 2013
~0.8PB -> a data center(!)
Gaussian trap
![Page 16: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/16.jpg)
CMU SCS
(c) 2013, C. Faloutsos 16
Solution# S.2: Eigen Exponent E
• A2: power law in the eigenvalues of the adjacency matrix
E = -0.48
Exponent = slope
Eigenvalue
Rank of decreasing eigenvalue
May 2001
DSBDE, July 2013
A x = λ x
![Page 17: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/17.jpg)
CMU SCS
(c) 2013, C. Faloutsos 17
Roadmap
• Introduction – Motivation • Problem#1: Patterns in graphs
– Static graphs • degree, diameter, eigen, • Triangles
– Time evolving graphs • Problem#2: Tools
DSBDE, July 2013
![Page 18: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/18.jpg)
CMU SCS
(c) 2013, C. Faloutsos 18
Solution# S.3: Triangle ‘Laws’
• Real social networks have a lot of triangles
DSBDE, July 2013
![Page 19: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/19.jpg)
CMU SCS
(c) 2013, C. Faloutsos 19
Solution# S.3: Triangle ‘Laws’
• Real social networks have a lot of triangles – Friends of friends are friends
• Any patterns? – 2x the friends, 2x the triangles ?
DSBDE, July 2013
![Page 20: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/20.jpg)
CMU SCS
(c) 2013, C. Faloutsos 20
Triangle Law: #S.3 [Tsourakakis ICDM 2008]
SN Reuters
Epinions X-axis: degree Y-axis: mean # triangles n friends -> ~n1.6 triangles
DSBDE, July 2013
![Page 21: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/21.jpg)
CMU SCS
(c) 2013, C. Faloutsos 21
Triangle Law: Computations [Tsourakakis ICDM 2008]
But: triangles are expensive to compute (3-way join; several approx. algos) – O(dmax
2) Q: Can we do that quickly? A:
details
DSBDE, July 2013
![Page 22: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/22.jpg)
CMU SCS
(c) 2013, C. Faloutsos 22
Triangle Law: Computations [Tsourakakis ICDM 2008]
But: triangles are expensive to compute (3-way join; several approx. algos) – O(dmax
2) Q: Can we do that quickly? A: Yes!
#triangles = 1/6 Sum ( λi3 )
(and, because of skewness (S2) , we only need the top few eigenvalues! - O(E)
details
DSBDE, July 2013
A x = λ x
![Page 23: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/23.jpg)
CMU SCS
Triangle counting for large graphs?
Anomalous nodes in Twitter(~ 3 billion edges) [U Kang, Brendan Meeder, +, PAKDD’11]
23 DSBDE, July 2013 23 (c) 2013, C. Faloutsos
? ?
?
![Page 24: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/24.jpg)
CMU SCS
Triangle counting for large graphs?
Anomalous nodes in Twitter(~ 3 billion edges) [U Kang, Brendan Meeder, +, PAKDD’11]
24 DSBDE, July 2013 24 (c) 2013, C. Faloutsos
![Page 25: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/25.jpg)
CMU SCS
Triangle counting for large graphs?
Anomalous nodes in Twitter(~ 3 billion edges) [U Kang, Brendan Meeder, +, PAKDD’11]
25 DSBDE, July 2013 25 (c) 2013, C. Faloutsos
![Page 26: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/26.jpg)
CMU SCS
Triangle counting for large graphs?
Anomalous nodes in Twitter(~ 3 billion edges) [U Kang, Brendan Meeder, +, PAKDD’11]
26 DSBDE, July 2013 26 (c) 2013, C. Faloutsos
![Page 27: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/27.jpg)
CMU SCS
(c) 2013, C. Faloutsos 27
Roadmap
• Introduction – Motivation • Part#1: Patterns in graphs
– Static graphs • Power law degrees; eigenvalues; triangles • Anti-pattern: NO good cuts!
– Time-evolving graphs • …. • Conclusions
DSBDE, July 2013
![Page 28: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/28.jpg)
CMU SCS
DSBDE, July 2013 (c) 2013, C. Faloutsos 28
Background: Graph cut problem
• Given a graph, and k • Break it into k (disjoint) communities
![Page 29: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/29.jpg)
CMU SCS
DSBDE, July 2013 (c) 2013, C. Faloutsos 29
Graph cut problem
• Given a graph, and k • Break it into k (disjoint) communities • (assume: block diagonal = ‘cavemen’ graph)
k = 2
![Page 30: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/30.jpg)
CMU SCS
Many algo’s for graph partitioning • METIS [Karypis, Kumar +] • 2nd eigenvector of Laplacian • Modularity-based [Girwan+Newman] • Max flow [Flake+] • … • … • …
DSBDE, July 2013 (c) 2013, C. Faloutsos 30
![Page 31: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/31.jpg)
CMU SCS
DSBDE, July 2013 (c) 2013, C. Faloutsos 31
Strange behavior of min cuts
• Subtle details: next – Preliminaries: min-cut plots of ‘usual’ graphs
NetMine: New Mining Tools for Large Graphs, by D. Chakrabarti, Y. Zhan, D. Blandford, C. Faloutsos and G. Blelloch, in the SDM 2004 Workshop on Link Analysis, Counter-terrorism and Privacy
Statistical Properties of Community Structure in Large Social and Information Networks, J. Leskovec, K. Lang, A. Dasgupta, M. Mahoney. WWW 2008.
![Page 32: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/32.jpg)
CMU SCS
DSBDE, July 2013 (c) 2013, C. Faloutsos 32
“Min-cut” plot • Do min-cuts recursively.
log (# edges)
log (mincut-size / #edges)
N nodes
Mincut size = sqrt(N)
![Page 33: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/33.jpg)
CMU SCS
DSBDE, July 2013 (c) 2013, C. Faloutsos 33
“Min-cut” plot • Do min-cuts recursively.
log (# edges)
log (mincut-size / #edges)
N nodes
New min-cut
![Page 34: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/34.jpg)
CMU SCS
DSBDE, July 2013 (c) 2013, C. Faloutsos 34
“Min-cut” plot • Do min-cuts recursively.
log (# edges)
log (mincut-size / #edges)
N nodes
New min-cut
Slope = -0.5
Better cut
![Page 35: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/35.jpg)
CMU SCS
DSBDE, July 2013 (c) 2013, C. Faloutsos 35
“Min-cut” plot
log (# edges)
log (mincut-size / #edges)
Slope = -1/d
For a d-dimensional grid, the slope is -1/d
log (# edges)
log (mincut-size / #edges)
For a random graph
(and clique),
the slope is 0
![Page 36: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/36.jpg)
CMU SCS
DSBDE, July 2013 (c) 2013, C. Faloutsos 36
Experiments • Datasets:
– Google Web Graph: 916,428 nodes and 5,105,039 edges
– Lucent Router Graph: Undirected graph of network routers from www.isi.edu/scan/mercator/maps.html; 112,969 nodes and 181,639 edges
– User Website Clickstream Graph: 222,704 nodes and 952,580 edges
NetMine: New Mining Tools for Large Graphs, by D. Chakrabarti, Y. Zhan, D. Blandford, C. Faloutsos and G. Blelloch, in the SDM 2004 Workshop on Link Analysis, Counter-terrorism and Privacy
![Page 37: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/37.jpg)
CMU SCS
DSBDE, July 2013 (c) 2013, C. Faloutsos 37
“Min-cut” plot • What does it look like for a real-world
graph?
log (# edges)
log (mincut-size / #edges)
?
![Page 38: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/38.jpg)
CMU SCS
DSBDE, July 2013 (c) 2013, C. Faloutsos 38
Experiments • Used the METIS algorithm [Karypis, Kumar,
1995]
-9-8-7-6-5-4-3-2-100510152025log(mincut/edges)log(edges)google-averaged"Lip"
log (# edges)
log
(min
cut-s
ize
/ #ed
ges)
• Google Web graph
• Values along the y-axis are averaged
• “lip” for large # edges
• Slope of -0.4, corresponds to a 2.5-dimensional grid!
Slope~ -0.4
log (# edges)
log (mincut-size / #edges)
![Page 39: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/39.jpg)
CMU SCS
DSBDE, July 2013 (c) 2013, C. Faloutsos 39
Experiments • Used the METIS algorithm [Karypis, Kumar,
1995]
-9-8-7-6-5-4-3-2-100510152025log(mincut/edges)log(edges)google-averaged"Lip"
log (# edges)
log
(min
cut-s
ize
/ #ed
ges)
• Google Web graph
• Values along the y-axis are averaged
• “lip” for large # edges
• Slope of -0.4, corresponds to a 2.5-dimensional grid!
Slope~ -0.4
log (# edges)
log (mincut-size / #edges)
Better cut
![Page 40: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/40.jpg)
CMU SCS
DSBDE, July 2013 (c) 2013, C. Faloutsos 40
Experiments • Same results for other graphs too…
-8-7-6-5-4-3-2-10024681012141618log(mincut/edges)log(edges)lucent-averaged
-4.5-4-3.5-3-2.5-2-1.5-1-0.5002468101214161820log(mincut/edges)log(edges)clickstream-averaged
log (# edges) log (# edges)
log
(min
cut-s
ize
/ #ed
ges)
log
(min
cut-s
ize
/ #ed
ges)
Lucent Router graph Clickstream graph
Slope~ -0.57 Slope~ -0.45
![Page 41: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/41.jpg)
CMU SCS
Why no good cuts? • Answer: self-similarity (few foils later)
DSBDE, July 2013 (c) 2013, C. Faloutsos 41
![Page 42: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/42.jpg)
CMU SCS
(c) 2013, C. Faloutsos 42
Roadmap
• Introduction – Motivation • Part#1: Patterns in graphs
– Static graphs – Time-evolving graphs – Why so many power-laws?
• Part#2: Cascade analysis • Conclusions
DSBDE, July 2013
![Page 43: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/43.jpg)
CMU SCS
(c) 2013, C. Faloutsos 43
Problem: Time evolution • with Jure Leskovec (CMU ->
Stanford)
• and Jon Kleinberg (Cornell – sabb. @ CMU)
DSBDE, July 2013
Jure Leskovec, Jon Kleinberg and Christos Faloutsos: Graphs over Time: Densification Laws, Shrinking Diameters and Possible Explanations, KDD 2005
![Page 44: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/44.jpg)
CMU SCS
(c) 2013, C. Faloutsos 44
T.1 Evolution of the Diameter • Prior work on Power Law graphs hints
at slowly growing diameter: – [diameter ~ O( N1/3)] – diameter ~ O(log N) – diameter ~ O(log log N)
• What is happening in real data?
DSBDE, July 2013
diameter
![Page 45: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/45.jpg)
CMU SCS
(c) 2013, C. Faloutsos 45
T.1 Evolution of the Diameter • Prior work on Power Law graphs hints
at slowly growing diameter: – [diameter ~ O( N1/3)] – diameter ~ O(log N) – diameter ~ O(log log N)
• What is happening in real data? • Diameter shrinks over time
DSBDE, July 2013
![Page 46: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/46.jpg)
CMU SCS
(c) 2013, C. Faloutsos 46
T.1 Diameter – “Patents”
• Patent citation network
• 25 years of data • @1999
– 2.9 M nodes – 16.5 M edges
time [years]
diameter
DSBDE, July 2013
![Page 47: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/47.jpg)
CMU SCS
(c) 2013, C. Faloutsos 47
T.2 Temporal Evolution of the Graphs
• N(t) … nodes at time t • E(t) … edges at time t • Suppose that
N(t+1) = 2 * N(t) • Q: what is your guess for
E(t+1) =? 2 * E(t)
DSBDE, July 2013
Say, k friends on average
k
![Page 48: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/48.jpg)
CMU SCS
(c) 2013, C. Faloutsos 48
T.2 Temporal Evolution of the Graphs
• N(t) … nodes at time t • E(t) … edges at time t • Suppose that
N(t+1) = 2 * N(t) • Q: what is your guess for
E(t+1) =? 2 * E(t) • A: over-doubled! ~ 3x
– But obeying the ``Densification Power Law’’ DSBDE, July 2013
Say, k friends on average
Gaussian trap
![Page 49: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/49.jpg)
CMU SCS
(c) 2013, C. Faloutsos 49
T.2 Temporal Evolution of the Graphs
• N(t) … nodes at time t • E(t) … edges at time t • Suppose that
N(t+1) = 2 * N(t) • Q: what is your guess for
E(t+1) =? 2 * E(t) • A: over-doubled! ~ 3x
– But obeying the ``Densification Power Law’’ DSBDE, July 2013
Say, k friends on average
Gaussian trap
✗ ✔ log
log
lin
lin
![Page 50: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/50.jpg)
CMU SCS
(c) 2013, C. Faloutsos 50
T.2 Densification – Patent Citations
• Citations among patents granted
• @1999 – 2.9 M nodes – 16.5 M edges
• Each year is a datapoint
N(t)
E(t)
1.66
DSBDE, July 2013
![Page 51: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/51.jpg)
CMU SCS
MORE Graph Patterns
DSBDE, July 2013 (c) 2013, C. Faloutsos 51 RTG: A Recursive Realistic Graph Generator using Random Typing Leman Akoglu and Christos Faloutsos. PKDD’09.
![Page 52: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/52.jpg)
CMU SCS
MORE Graph Patterns
DSBDE, July 2013 (c) 2013, C. Faloutsos 52
✔
✔ ✔ ✔
✔
RTG: A Recursive Realistic Graph Generator using Random Typing Leman Akoglu and Christos Faloutsos. PKDD’09.
![Page 53: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/53.jpg)
CMU SCS
MORE Graph Patterns
DSBDE, July 2013 (c) 2013, C. Faloutsos 53
• Mary McGlohon, Leman Akoglu, Christos Faloutsos. Statistical Properties of Social Networks. in "Social Network Data Analytics” (Ed.: Charu Aggarwal)
• Deepayan Chakrabarti and Christos Faloutsos, Graph Mining: Laws, Tools, and Case Studies Oct. 2012, Morgan Claypool.
![Page 54: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/54.jpg)
CMU SCS
(c) 2013, C. Faloutsos 54
Roadmap
• Introduction – Motivation • Part#1: Patterns in graphs
– … – Why so many power-laws? – Why no ‘good cuts’?
• Part#2: Cascade analysis • Conclusions
DSBDE, July 2013
![Page 55: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/55.jpg)
CMU SCS
2 Questions, one answer • Q1: why so many power laws • Q2: why no ‘good cuts’?
DSBDE, July 2013 (c) 2013, C. Faloutsos 55
![Page 56: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/56.jpg)
CMU SCS
2 Questions, one answer • Q1: why so many power laws • Q2: why no ‘good cuts’? • A: Self-similarity = fractals = ‘RMAT’ ~
‘Kronecker graphs’
DSBDE, July 2013 (c) 2013, C. Faloutsos 56
possible
![Page 57: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/57.jpg)
CMU SCS
20’’ intro to fractals • Remove the middle triangle; repeat • -> Sierpinski triangle • (Bonus question - dimensionality?
– >1 (inf. perimeter – (4/3)∞ ) – <2 (zero area – (3/4) ∞ )
DSBDE, July 2013 (c) 2013, C. Faloutsos 57
…
![Page 58: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/58.jpg)
CMU SCS
20’’ intro to fractals
DSBDE, July 2013 (c) 2013, C. Faloutsos 58
Self-similarity -> no char. scale -> power laws, eg: 2x the radius, 3x the #neighbors nn(r) nn(r) = C r log3/log2
![Page 59: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/59.jpg)
CMU SCS
20’’ intro to fractals
DSBDE, July 2013 (c) 2013, C. Faloutsos 59
Self-similarity -> no char. scale -> power laws, eg: 2x the radius, 3x the #neighbors nn(r) nn(r) = C r log3/log2
![Page 60: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/60.jpg)
CMU SCS
20’’ intro to fractals
DSBDE, July 2013 (c) 2013, C. Faloutsos 60
Self-similarity -> no char. scale -> power laws, eg: 2x the radius, 3x the #neighbors nn = C r log3/log2
N(t)
E(t)
1.66
Reminder: Densification P.L. (2x nodes, ~3x edges)
![Page 61: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/61.jpg)
CMU SCS
20’’ intro to fractals
DSBDE, July 2013 (c) 2013, C. Faloutsos 61
Self-similarity -> no char. scale -> power laws, eg: 2x the radius, 3x the #neighbors nn = C r log3/log2
2x the radius, 4x neighbors
nn = C r log4/log2 = C r 2
![Page 62: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/62.jpg)
CMU SCS
20’’ intro to fractals
DSBDE, July 2013 (c) 2013, C. Faloutsos 62
Self-similarity -> no char. scale -> power laws, eg: 2x the radius, 3x the #neighbors nn = C r log3/log2
2x the radius, 4x neighbors
nn = C r log4/log2 = C r 2
Fractal dim.
=1.58
![Page 63: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/63.jpg)
CMU SCS
20’’ intro to fractals
DSBDE, July 2013 (c) 2013, C. Faloutsos 63
Self-similarity -> no char. scale -> power laws, eg: 2x the radius, 3x the #neighbors nn = C r log3/log2
2x the radius, 4x neighbors
nn = C r log4/log2 = C r 2
Fractal dim.
![Page 64: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/64.jpg)
CMU SCS How does self-similarity help in
graphs? • A: RMAT/Kronecker generators
– With self-similarity, we get all power-laws, automatically,
– And small/shrinking diameter – And `no good cuts’
DSBDE, July 2013 (c) 2013, C. Faloutsos 64
R-MAT: A Recursive Model for Graph Mining, by D. Chakrabarti, Y. Zhan and C. Faloutsos, SDM 2004, Orlando, Florida, USA Realistic, Mathematically Tractable Graph Generation and Evolution, Using Kronecker Multiplication, by J. Leskovec, D. Chakrabarti, J. Kleinberg, and C. Faloutsos, in PKDD 2005, Porto, Portugal
![Page 65: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/65.jpg)
CMU SCS
DSBDE, July 2013 (c) 2013, C. Faloutsos 65
Graph gen.: Problem dfn • Given a growing graph with count of nodes N1,
N2, … • Generate a realistic sequence of graphs that will
obey all the patterns – Static Patterns
S1 Power Law Degree Distribution S2 Power Law eigenvalue and eigenvector distribution Small Diameter
– Dynamic Patterns T2 Growth Power Law (2x nodes; 3x edges) T1 Shrinking/Stabilizing Diameters
![Page 66: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/66.jpg)
CMU SCS
DSBDE, July 2013 (c) 2013, C. Faloutsos 66 Adjacency matrix
Kronecker Graphs
Intermediate stage
Adjacency matrix
![Page 67: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/67.jpg)
CMU SCS
DSBDE, July 2013 (c) 2013, C. Faloutsos 67 Adjacency matrix
Kronecker Graphs
Intermediate stage
Adjacency matrix
![Page 68: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/68.jpg)
CMU SCS
DSBDE, July 2013 (c) 2013, C. Faloutsos 68 Adjacency matrix
Kronecker Graphs
Intermediate stage
Adjacency matrix
![Page 69: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/69.jpg)
CMU SCS
DSBDE, July 2013 (c) 2013, C. Faloutsos 69
Kronecker Graphs • Continuing multiplying with G1 we obtain G4 and
so on …
G4 adjacency matrix
![Page 70: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/70.jpg)
CMU SCS
DSBDE, July 2013 (c) 2013, C. Faloutsos 70
Kronecker Graphs • Continuing multiplying with G1 we obtain G4 and
so on …
G4 adjacency matrix
![Page 71: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/71.jpg)
CMU SCS
DSBDE, July 2013 (c) 2013, C. Faloutsos 71
Kronecker Graphs • Continuing multiplying with G1 we obtain G4 and
so on …
G4 adjacency matrix
![Page 72: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/72.jpg)
CMU SCS
DSBDE, July 2013 (c) 2013, C. Faloutsos 72
Kronecker Graphs • Continuing multiplying with G1 we obtain G4 and
so on …
G4 adjacency matrix
Holes within holes; Communities
within communities
![Page 73: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/73.jpg)
CMU SCS
DSBDE, July 2013 (c) 2013, C. Faloutsos 73
Properties:
• We can PROVE that – Degree distribution is multinomial ~ power law – Diameter: constant – Eigenvalue distribution: multinomial – First eigenvector: multinomial
new
Self-similarity -> power laws
![Page 74: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/74.jpg)
CMU SCS
DSBDE, July 2013 (c) 2013, C. Faloutsos 74
Problem Definition • Given a growing graph with nodes N1, N2, … • Generate a realistic sequence of graphs that will obey all
the patterns – Static Patterns
Power Law Degree Distribution Power Law eigenvalue and eigenvector distribution Small Diameter
– Dynamic Patterns Growth Power Law Shrinking/Stabilizing Diameters
• First generator for which we can prove all these properties
![Page 75: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/75.jpg)
CMU SCS
Impact: Graph500 • Based on RMAT (= 2x2 Kronecker) • Standard for graph benchmarks • http://www.graph500.org/ • Competitions 2x year, with all major
entities: LLNL, Argonne, ITC-U. Tokyo, Riken, ORNL, Sandia, PSC, …
DSBDE, July 2013 (c) 2013, C. Faloutsos 75
R-MAT: A Recursive Model for Graph Mining, by D. Chakrabarti, Y. Zhan and C. Faloutsos, SDM 2004, Orlando, Florida, USA
To iterate is human, to recurse is devine
![Page 76: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/76.jpg)
CMU SCS
(c) 2013, C. Faloutsos 76
Roadmap
• Introduction – Motivation • Part#1: Patterns in graphs
– … – Q1: Why so many power-laws? – Q2: Why no ‘good cuts’?
• Part#2: Cascade analysis • Conclusions
DSBDE, July 2013
A: real graphs -> self similar -> power laws
![Page 77: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/77.jpg)
CMU SCS
Q2: Why ‘no good cuts’? • A: self-similarity
– Communities within communities within communities …
DSBDE, July 2013 (c) 2013, C. Faloutsos 77
![Page 78: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/78.jpg)
CMU SCS
DSBDE, July 2013 (c) 2013, C. Faloutsos 78
Kronecker Product – a Graph • Continuing multiplying with G1 we obtain G4 and
so on …
G4 adjacency matrix
REMINDER
![Page 79: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/79.jpg)
CMU SCS
DSBDE, July 2013 (c) 2013, C. Faloutsos 79
Kronecker Product – a Graph • Continuing multiplying with G1 we obtain G4 and
so on …
G4 adjacency matrix
REMINDER
Communities within communities within communities …
‘Linux users’
‘Mac users’
‘win users’
![Page 80: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/80.jpg)
CMU SCS
DSBDE, July 2013 (c) 2013, C. Faloutsos 80
Kronecker Product – a Graph • Continuing multiplying with G1 we obtain G4 and
so on …
G4 adjacency matrix
REMINDER
Communities within communities within communities …
How many Communities? 3? 9? 27?
![Page 81: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/81.jpg)
CMU SCS
DSBDE, July 2013 (c) 2013, C. Faloutsos 81
Kronecker Product – a Graph • Continuing multiplying with G1 we obtain G4 and
so on …
G4 adjacency matrix
REMINDER
Communities within communities within communities …
How many Communities? 3? 9? 27?
A: one – but not a typical, block-like community…
![Page 82: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/82.jpg)
CMU SCS
DSBDE, July 2013 (c) 2013, C. Faloutsos 82
Communities? (Gaussian) Clusters?
Piece-wise flat parts?
age
# songs
![Page 83: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/83.jpg)
CMU SCS
DSBDE, July 2013 (c) 2013, C. Faloutsos 83
Wrong questions to ask!
age
# songs
![Page 84: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/84.jpg)
CMU SCS
Summary of Part#1 • *many* patterns in real graphs
– Small & shrinking diameters – Power-laws everywhere – Gaussian trap – ‘no good cuts’
• Self-similarity (RMAT/Kronecker): good model
DSBDE, July 2013 (c) 2013, C. Faloutsos 84
![Page 85: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/85.jpg)
CMU SCS
DSBDE, July 2013 (c) 2013, C. Faloutsos 85
Part 2: Cascades &
Immunization
![Page 86: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/86.jpg)
CMU SCS
Why do we care? • Information Diffusion • Viral Marketing • Epidemiology and Public Health • Cyber Security • Human mobility • Games and Virtual Worlds • Ecology • ........
(c) 2013, C. Faloutsos 86 DSBDE, July 2013
![Page 87: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/87.jpg)
CMU SCS
(c) 2013, C. Faloutsos 87
Roadmap
• Introduction – Motivation • Part#1: Patterns in graphs • Part#2: Cascade analysis
– (Fractional) Immunization – Epidemic thresholds
• Conclusions
DSBDE, July 2013
![Page 88: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/88.jpg)
CMU SCS
Fractional Immunization of Networks B. Aditya Prakash, Lada Adamic, Theodore Iwashyna (M.D.), Hanghang Tong,
Christos Faloutsos
SDM 2013, Austin, TX
(c) 2013, C. Faloutsos 88 DSBDE, July 2013
![Page 89: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/89.jpg)
CMU SCS
Whom to immunize? • Dynamical Processes over networks
• Each circle is a hospital • ~3,000 hospitals • More than 30,000 pa4ents transferred
[US-‐MEDICARE NETWORK 2005]
Problem: Given k units of disinfectant, whom to immunize?
(c) 2013, C. Faloutsos 89 DSBDE, July 2013
![Page 90: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/90.jpg)
CMU SCS
Whom to immunize?
CURRENT PRACTICE OUR METHOD
[US-‐MEDICARE NETWORK 2005]
~6x fewer!
(c) 2013, C. Faloutsos 90 DSBDE, July 2013
Hospital-acquired inf. : 99K+ lives, $5B+ per year
![Page 91: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/91.jpg)
CMU SCS
Fractional Asymmetric Immunization
Hospital Another Hospital
Drug-‐resistant Bacteria (like XDR-‐TB)
(c) 2013, C. Faloutsos 91 DSBDE, July 2013
![Page 92: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/92.jpg)
CMU SCS
Fractional Asymmetric Immunization
Hospital Another Hospital
(c) 2013, C. Faloutsos 92 DSBDE, July 2013
![Page 93: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/93.jpg)
CMU SCS
Fractional Asymmetric Immunization
Hospital Another Hospital
(c) 2013, C. Faloutsos 93 DSBDE, July 2013
![Page 94: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/94.jpg)
CMU SCS
Fractional Asymmetric Immunization
Hospital Another Hospital
Problem: Given k units of disinfectant, distribute them to maximize hospitals saved
(c) 2013, C. Faloutsos 94 DSBDE, July 2013
![Page 95: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/95.jpg)
CMU SCS
Fractional Asymmetric Immunization
Hospital Another Hospital
Problem: Given k units of disinfectant, distribute them to maximize hospitals saved @ 365 days
(c) 2013, C. Faloutsos 95 DSBDE, July 2013
![Page 96: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/96.jpg)
CMU SCS
Straightforward solution: Simulation: 1. Distribute resources 2. ‘infect’ a few nodes 3. Simulate evolution of spreading
– (10x, take avg) 4. Tweak, and repeat step 1
DSBDE, July 2013 (c) 2013, C. Faloutsos 96
![Page 97: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/97.jpg)
CMU SCS
Straightforward solution: Simulation: 1. Distribute resources 2. ‘infect’ a few nodes 3. Simulate evolution of spreading
– (10x, take avg) 4. Tweak, and repeat step 1
DSBDE, July 2013 (c) 2013, C. Faloutsos 97
![Page 98: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/98.jpg)
CMU SCS
Straightforward solution: Simulation: 1. Distribute resources 2. ‘infect’ a few nodes 3. Simulate evolution of spreading
– (10x, take avg) 4. Tweak, and repeat step 1
DSBDE, July 2013 (c) 2013, C. Faloutsos 98
![Page 99: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/99.jpg)
CMU SCS
Straightforward solution: Simulation: 1. Distribute resources 2. ‘infect’ a few nodes 3. Simulate evolution of spreading
– (10x, take avg) 4. Tweak, and repeat step 1
DSBDE, July 2013 (c) 2013, C. Faloutsos 99
![Page 100: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/100.jpg)
CMU SCS
Running Time
Simula4ons SMART-‐ALLOC
> 1 week Wall-‐Clock
Time ≈
14 secs
> 30,000x speed-‐up!
be?er
(c) 2013, C. Faloutsos 100 DSBDE, July 2013
![Page 101: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/101.jpg)
CMU SCS
Experiments
K = 120
be?er
(c) 2013, C. Faloutsos 101 DSBDE, July 2013
# epochs
# infected uniform
SMART-ALLOC
![Page 102: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/102.jpg)
CMU SCS
What is the ‘silver bullet’? A: Try to decrease connectivity of graph
Q: how to measure connectivity? – Avg degree? Max degree? – Std degree / avg degree ? – Diameter? – Modularity? – ‘Conductance’ (~min cut size)? – Some combination of above?
DSBDE, July 2013 (c) 2013, C. Faloutsos 102
14 secs
> 30,000x speed-‐up!
![Page 103: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/103.jpg)
CMU SCS
What is the ‘silver bullet’? A: Try to decrease connectivity of graph
Q: how to measure connectivity? A: first eigenvalue of adjacency matrix
Q1: why?? (Q2: dfn & intuition of eigenvalue ? )
DSBDE, July 2013 (c) 2013, C. Faloutsos 103
Avg degree Max degree Diameter Modularity ‘Conductance’
![Page 104: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/104.jpg)
CMU SCS
Why eigenvalue? A1: ‘G2’ theorem and ‘eigen-drop’:
• For (almost) any type of virus • For any network • -> no epidemic, if small-enough first
eigenvalue (λ1 ) of adjacency matrix
• Heuristic: for immunization, try to min λ1
• The smaller λ1, the closer to extinction. DSBDE, July 2013 (c) 2013, C. Faloutsos 104
Threshold Conditions for Arbitrary Cascade Models on Arbitrary Networks, B. Aditya Prakash, Deepayan Chakrabarti, Michalis Faloutsos, Nicholas Valler, Christos Faloutsos, ICDM 2011, Vancouver, Canada
![Page 105: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/105.jpg)
CMU SCS
Why eigenvalue? A1: ‘G2’ theorem and ‘eigen-drop’:
• For (almost) any type of virus • For any network • -> no epidemic, if small-enough first
eigenvalue (λ1 ) of adjacency matrix
• Heuristic: for immunization, try to min λ1
• The smaller λ1, the closer to extinction. DSBDE, July 2013 (c) 2013, C. Faloutsos 105
![Page 106: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/106.jpg)
CMU SCS
Threshold Conditions for Arbitrary Cascade Models on Arbitrary Networks B. Aditya Prakash, Deepayan Chakrabarti, Michalis Faloutsos, Nicholas Valler, Christos Faloutsos IEEE ICDM 2011, Vancouver
extended version, in arxiv http://arxiv.org/abs/1004.0060
G2 theorem
~10 pages proof
![Page 107: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/107.jpg)
CMU SCS
Our thresholds for some models • s = effective strength • s < 1 : below threshold
(c) 2013, C. Faloutsos 107 DSBDE, July 2013
Models Effective Strength (s)
Threshold (tipping point)
SIS, SIR, SIRS, SEIR s = λ .
s = 1 SIV, SEIV s = λ .
(H.I.V.) s = λ .
![Page 108: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/108.jpg)
CMU SCS
Our thresholds for some models • s = effective strength • s < 1 : below threshold
(c) 2013, C. Faloutsos 108 DSBDE, July 2013
Models Effective Strength (s)
Threshold (tipping point)
SIS, SIR, SIRS, SEIR s = λ .
s = 1 SIV, SEIV s = λ .
(H.I.V.) s = λ .
No immunity
Temp. immunity
w/ incubation
![Page 109: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/109.jpg)
CMU SCS
(c) 2013, C. Faloutsos 109
Roadmap
• Introduction – Motivation • Part#1: Patterns in graphs • Part#2: Cascade analysis
– (Fractional) Immunization – intuition behind λ1
• Conclusions
DSBDE, July 2013
![Page 110: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/110.jpg)
CMU SCS
Intuition for λ
“Official” definitions: • Let A be the adjacency
matrix. Then λ is the root with the largest magnitude of the characteristic polynomial of A [det(A – xI)].
• Also: A x = λ x
Neither gives much intuition!
“Un-official” Intuition • For ‘homogeneous’
graphs, λ == degree
• λ ~ avg degree – done right, for skewed
degree distributions
(c) 2013, C. Faloutsos 110 DSBDE, July 2013
![Page 111: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/111.jpg)
CMU SCS
Largest Eigenvalue (λ)
λ ≈ 2 λ = N λ = N-‐1
N = 1000 nodes λ ≈ 2 λ= 31.67 λ= 999
beaer connec4vity higher λ �
(c) 2013, C. Faloutsos 111 DSBDE, July 2013
![Page 112: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/112.jpg)
CMU SCS
Largest Eigenvalue (λ)
λ ≈ 2 λ = N λ = N-‐1
N = 1000 nodes λ ≈ 2 λ= 31.67 λ= 999
beaer connec4vity higher λ �
(c) 2013, C. Faloutsos 112 DSBDE, July 2013
![Page 113: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/113.jpg)
CMU SCS
Examples: Simulations – SIR (mumps) Fr
actio
n of
Infe
ctio
ns
Foot
prin
t Effective Strength Time ticks
(c) 2013, C. Faloutsos 113 DSBDE, July 2013
(a) Infection profile (b) “Take-off” plot PORTLAND graph: syntheGc populaGon,
31 million links, 6 million nodes
![Page 114: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/114.jpg)
CMU SCS Examples: Simulations – SIRS
(pertusis)
Frac
tion
of In
fect
ions
Foot
prin
t Effective Strength Time ticks
(c) 2013, C. Faloutsos 114 DSBDE, July 2013
(a) Infection profile (b) “Take-off” plot PORTLAND graph: syntheGc populaGon,
31 million links, 6 million nodes
![Page 115: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/115.jpg)
CMU SCS
Immunization - conclusion In (almost any) immunization setting, • Allocate resources, such that to • Minimize λ1
• (regardless of virus specifics)
• Conversely, in a market penetration setting – Allocate resources to – Maximize λ1
DSBDE, July 2013 (c) 2013, C. Faloutsos 115
![Page 116: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/116.jpg)
CMU SCS
(c) 2013, C. Faloutsos 116
Roadmap
• Introduction – Motivation • Part#1: Patterns in graphs • Part#2: Cascade analysis
– (Fractional) Immunization – Epidemic thresholds
• What next? • Acks & Conclusions • [Tools: ebay fraud; tensors; spikes] DSBDE, July 2013
![Page 117: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/117.jpg)
CMU SCS Challenge #1: ‘Connectome’ –
brain wiring
DSBDE, July 2013 (c) 2013, C. Faloutsos 117
• Which neurons get activated by ‘bee’ • How wiring evolves • Modeling epilepsy
50 100 150 200 250
50
100
150
200
250
3000
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0.04
N. Sidiropoulos George Karypis
V. Papalexakis
Tom Mitchell
![Page 118: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/118.jpg)
CMU SCS Challenge#2: Time evolving
networks / tensors • Periodicities? Burstiness? • What is ‘typical’ behavior of a node, over time • Heterogeneous graphs (= nodes w/ attributes)
DSBDE, July 2013 (c) 2013, C. Faloutsos 118
…
![Page 119: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/119.jpg)
CMU SCS
(c) 2013, C. Faloutsos 119
Roadmap
• Introduction – Motivation • Part#1: Patterns in graphs • Part#2: Cascade analysis
– (Fractional) Immunization – Epidemic thresholds
• Acks & Conclusions • [Tools: ebay fraud; tensors; spikes]
DSBDE, July 2013
Off line
![Page 120: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/120.jpg)
CMU SCS
(c) 2013, C. Faloutsos 120
Thanks
DSBDE, July 2013
Thanks to: NSF IIS-0705359, IIS-0534205, CTA-INARC; Yahoo (M45), LLNL, IBM, SPRINT, Google, INTEL, HP, iLab
![Page 121: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/121.jpg)
CMU SCS
(c) 2013, C. Faloutsos 121
Thanks
DSBDE, July 2013
Thanks to: NSF IIS-0705359, IIS-0534205, CTA-INARC; Yahoo (M45), LLNL, IBM, SPRINT, Google, INTEL, HP, iLab
Disclaimer: All opinions are mine; not necessarily reflecting the opinions of the funding agencies
![Page 122: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/122.jpg)
CMU SCS
(c) 2013, C. Faloutsos 122
Project info: PEGASUS
DSBDE, July 2013
www.cs.cmu.edu/~pegasus Results on large graphs: with Pegasus +
hadoop + M45 Apache license Code, papers, manual, video
Prof. U Kang Prof. Polo Chau
![Page 123: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/123.jpg)
CMU SCS
(c) 2013, C. Faloutsos 123
Cast
Akoglu, Leman
Chau, Polo
Kang, U
McGlohon, Mary
Tong, Hanghang
Prakash, Aditya
DSBDE, July 2013
Koutra, Danai
Beutel, Alex
Papalexakis, Vagelis
![Page 124: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/124.jpg)
CMU SCS
(c) 2013, C. Faloutsos 124
CONCLUSION#1 – Big data
• Large datasets reveal patterns/outliers that are invisible otherwise
DSBDE, July 2013
![Page 125: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/125.jpg)
CMU SCS
(c) 2013, C. Faloutsos 125
CONCLUSION#2 – self-similarity
• powerful tool / viewpoint – Power laws; shrinking diameters
– Gaussian trap (eg., F.O.F.)
– ‘no good cuts’
– RMAT – graph500 generator
DSBDE, July 2013
![Page 126: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/126.jpg)
CMU SCS
(c) 2013, C. Faloutsos 126
CONCLUSION#3 – eigen-drop
• Cascades & immunization: G2 theorem & eigenvalue
DSBDE, July 2013
CURRENT PRACTICE OUR METHOD
[US-‐MEDICARE NETWORK 2005]
~6x fewer!
14 secs
> 30,000x speed-‐up!
![Page 127: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/127.jpg)
CMU SCS
(c) 2013, C. Faloutsos 127
References • D. Chakrabarti, C. Faloutsos: Graph Mining – Laws,
Tools and Case Studies, Morgan Claypool 2012 • http://www.morganclaypool.com/doi/abs/10.2200/
S00449ED1V01Y201209DMK006
DSBDE, July 2013
![Page 128: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/128.jpg)
CMU SCS
TAKE HOME MESSAGE: Cross-disciplinarity
DSBDE, July 2013 (c) 2013, C. Faloutsos 128
![Page 129: Large Graph Mining - Patterns, Explanations and Cascade ...christos/TALKS/13-07-DSBDE/FOILS/faloutsos_dsbde_2013.pdflog(mincut/edges) lo g(ed g es) google-averaged "Lip" s) log (#](https://reader034.vdocuments.net/reader034/viewer/2022042911/5f421c324b43d37c481a54eb/html5/thumbnails/129.jpg)
CMU SCS
TAKE HOME MESSAGE: Cross-disciplinarity
DSBDE, July 2013 (c) 2013, C. Faloutsos 129
QUESTIONS?
www.cs.cmu.edu/~christos/TALKS/13-07-DSBDE/