Mobile Networks 2015.08.08 10:30 2015.08.08 10:48 2015.08.08 11:01 - - PowerPoint PPT Presentation

▶
mobile networks
SMART_READER_LITE
LIVE PREVIEW

Mobile Networks 2015.08.08 10:30 2015.08.08 10:48 2015.08.08 11:01 - - PowerPoint PPT Presentation

Dec 15, 2023 128 likes •561 views

CoupledLP: Link Prediction in Coupled Networks Yuxiao Dong # , Jing Zhang + , Jie Tang + , Nitesh V. Chawla # , Bai Wang* * Beijing University of Posts # University of Notre Dame + Tsinghua University and Telecommunications 1 Mobile Networks

slide-1
SLIDE 1

1

CoupledLP: Link Prediction in Coupled Networks

Yuxiao Dong#, Jing Zhang+, Jie Tang+, Nitesh V. Chawla#, Bai Wang*

#University of Notre Dame +Tsinghua University *Beijing University of Posts

and Telecommunications

slide-2
SLIDE 2

2

2015.08.08 10:30 2015.08.08 10:48 2015.08.08 11:01 2015.08.08 11:29 …… 2016.01.01 00:00 ……

Mobile Networks

slide-3
SLIDE 3

3

2015.08.08 10:30 2015.08.08 10:48 2015.08.08 11:01 2015.08.08 11:29 …… 2016.01.01 00:00 ……

Mobile Networks

slide-4
SLIDE 4

4

Disease-Gene Networks

1. K.I. Goh, M. E. Cusick, D. Valle, B. Childs, M. Vidal, and A.-L. Barabási. The human disease network. PNAS 2007. 2.

  • D. Davis, N. V. Chawla. Exploring and Exploiting Disease Interactions from Multi-Relational Gene and Phenotype Networks. PLoS One 2011.

3.

  • J. Menche, A. Sharma, M. Kitsak, S. D. Ghiassian, M. Vidal, J. Loscalzo, A.-L. Barabási. Uncovering disease-disease relationships through the incomplete interactome. Science 2015.

Disease network Cross network Gene network

slide-5
SLIDE 5

5

Coupled Networks

Given a source network GS = (VS, ES) and a target network GT = (VT, ET), they compose coupled networks if there exists a cross link eij with one node vi ∈ VS and the other node vj ∈ VT. The cross network GC = (VC, EC) is a bipartite network containing all the cross links in the coupled networks.

Source network Target network Cross network Coupled networks

slide-6
SLIDE 6

6

Coupled Link Prediction

Input ¡ Output ¡

Source network Cross network Target network

Given the source network GS and the cross network GC in coupled networks G = (GS, GT, GC), the task is to find a predictive function: f : (GS, GC) → YT where YT is the set of labels for the potential links in the target network GT.

slide-7
SLIDE 7

7

Challenges

Input ¡ Output ¡

Source network Cross network Target network

I n c

  • m

p l e t e n e s s H e t e r

  • g

e n e i t y A s y m m e t r y

slide-8
SLIDE 8

8

Related Work: Traditional Link Prediction

  • 1. D. Liben-Nowell and J. Kleinberg. The link prediction problem for social networks. CIKM’03.

B D C A E F G

t1

B D C A E F G Input Output

t2

slide-9
SLIDE 9

9

Related Work: Heterogeneous Link Prediction

Input

  • 1. Y. Sun, J. Han, C. C. Aggarwal, N. V. Chawla. Will Will This Happen? Relationship Prediction in Heterogeneous Information Networks. WSDM’12.

Output

slide-10
SLIDE 10

10

Related Work: Transfer Link Prediction

Input Output

1.

  • Y. Dong, J. Tang, S. Wu, J. Tian, N. V. Chawla. J. Rao, H. Cao. Link Prediction and Recommendation across Heterogeneous Networks. ICDM’12

2.

  • J. Tang, T. Lou, J. Kleinberg, S. Wu. Transfer Link Prediction across Heterogeneous Networks. TOIS 2015.

Source network Target network

slide-11
SLIDE 11

11

Related Work: Cross-Domain Link Prediction

Input Output

  • 1. J. Tang, S. Wu, J. Sun, H. Su. Cross-Domain Collaboration Recommendation. KDD’12.

Source network Cross network Target network

slide-12
SLIDE 12

12

Related Work: Anchor Link Prediction

Input Output

1.

  • X. Kong, J. Zhang, P. S. Yu. Inferring anchor links across multiple heterogeneous social networks. CIKM’13.

2.

  • Y. Zhang, J. Tang, Z. Yang, J. Pei, and P. S. Yu. COSNET: Connecting Heterogeneous Social Networks with Local and Global Consistency. KDD’15.

A network Self-linkage network B network

slide-13
SLIDE 13

13

Challenges

Input ¡ Output ¡

Source network Cross network Target network

I n c

  • m

p l e t e n e s s H e t e r

  • g

e n e i t y A s y m m e t r y

slide-14
SLIDE 14

14

CoupledLP Framework

  • 1. Implicit Target Network Construction
  • Solve Incompleteness
  • 2. Coupled Factor Graph Model
  • Solve Asymmetry
  • Solve Heterogeneity
slide-15
SLIDE 15

15

CoupledLP Framework

Input Output

Source network Cross network Target network 99% 80% 75% 87%

Implicit Target Network Construction

Incompleteness

  • 1. V. Leroy, B. B. Cambazoglu, and F. Bonchi. Cold start link prediction. In KDD ’10.
slide-16
SLIDE 16

16

CoupledLP: Implicit Target Network

v3

S

v1

T

v4

S

v2

T

v3

S

v1

T

v4

S

v2

T

v3

S

v1

T

v4

S

v2

T

  • 1. R. Guha, R. Kumar, P. Raghavan, and A. Tomkins. Propagation of trust and distrust. In WWW ’04

Atomic Propagations for constructing an implicit target network

Direct Coupling Co-citation

MM MMT MTM

+ +

top z%

slide-17
SLIDE 17

17

CoupledLP Framework

Input Output

Source network Cross network Target network 99% 80% 75% 87%

Coupled Factor Graph Asymmetry Heterogeneity

slide-18
SLIDE 18

18

Basic Idea

y12 y34 y23

v1 v2 v3 v4 v5 v6

1.

  • F. R. Kschischang, B. J. Frey, and H. andrea Loeliger. Factor graphs and the sum-product algorithm. In IEEE TOIT 2001.

2. Jie Tang, Tiancheng Lou, and Jon Kleinberg. Inferring social ties across heterogeneous networks. In WSDM '12

Input Network Factor Graph

slide-19
SLIDE 19

19

CoupledLP: Coupled Factor Graph

y13

f (v1, v3, y13)

y34 y06

Observations y13=? v1,v3 v3, v4 v0, v6 y06=?

f (v3, v4, y34) F (vs, v3, ys3)

y12

v1, v2

f (v1, v2, y12)

y12=1 v5, v6 v2, v3

y56 y23

y23=? y56=0

f (v2, v3, y23) f (v5, v6, y56) g(y12, y13) g(y13, y23) g(y06, y56)

v5 v2 v3 v0

Coupled Networks

v6 v1 v4

...

CoupledFG

g(y12, y23)

y34=?

slide-20
SLIDE 20

20

CoupledLP: Coupled Factor Graph

y13

f (v1, v3, y13)

y34 y06

Observations y13=? v1,v3 v3, v4 v0, v6 y06=?

f (v3, v4, y34) F (vs, v3, ys3)

y12

v1, v2

f (v1, v2, y12)

y12=1 v5, v6 v2, v3

y56 y23

y23=? y56=0

f (v2, v3, y23) f (v5, v6, y56) g(y12, y13) g(y13, y23) g(y06, y56)

v5 v2 v3 v0

Coupled Networks

v6 v1 v4

...

CoupledFG

g(y12, y23)

y34=?

model source and target network separately meta-path

Asymmetry

Attribute Factor

  • 1. Jie Tang, Tiancheng Lou, and Jon Kleinberg. Inferring social ties across heterogeneous networks. In WSDM '12.
slide-21
SLIDE 21

21

CoupledLP: Coupled Factor Graph

y13

f (v1, v3, y13)

y34 y06

Observations y13=? v1,v3 v3, v4 v0, v6 y06=?

f (v3, v4, y34) F (vs, v3, ys3)

y12

v1, v2

f (v1, v2, y12)

y12=1 v5, v6 v2, v3

y56 y23

y23=? y56=0

f (v2, v3, y23) f (v5, v6, y56) g(y12, y13) g(y13, y23) g(y06, y56)

v5 v2 v3 v0

Coupled Networks

v6 v1 v4

...

CoupledFG

g(y12, y23)

y34=?

model source and target network separately meta-path

Heterogeneity Asymmetry

Meta-path Factor

  • 1. Jie Tang, Tiancheng Lou, and Jon Kleinberg. Inferring social ties across heterogeneous networks. In WSDM '12.
slide-22
SLIDE 22

22

CoupledLP: Coupled Factor Graph

1.

  • Y. Sun, J. Han, C. C. Aggarwal, N. V. Chawla. Will Will This Happen? Relationship Prediction in Heterogeneous Information Networks. In WSDM’12.

meta-path

disease

express associate

gene gene disease disease gene gene

? ?

……

slide-23
SLIDE 23

23

v Factor Initialization: exponential-linear

CoupledLP: Coupled Factor Graph

v Objective Function:

model source & target network separately meta-path bridge source & target networks

slide-24
SLIDE 24

24

Learning: Gradient Decent method

  • 1. Jie Tang, Tiancheng Lou, and Jon Kleinberg. Inferring social ties across heterogeneous networks. In WSDM '12.

CoupledLP: Coupled Factor Graph

slide-25
SLIDE 25

25

CoupledLP Framework

  • 1. Implicit Target Network Construction
  • Solve Incompleteness
  • 2. Coupled Factor Graph Model
  • Solve Asymmetry
  • Solve Heterogeneity
slide-26
SLIDE 26

26

Experiments: Data

1.

  • D. Davis, N. V. Chawla. Exploring and Exploiting Disease Interactions from Multi-Relational Gene and Phenotype Networks. PLoS One 2011.

2.

  • N. Du, C. Faloutsos, B. Wang, and L. Akoglu. Large human communication networks: patterns and a utility-driven generator. In KDD ’09.

3.

  • Y. Dong, Y. Yang, J. Tang, Y. Yang, and N. V. Chawla. Inferring user demographics and social strategies in mobile social networks. In KDD’14.

k: average degree; cc: clustering coefficient; ac: associative coefficient

Healthcare Networks Disease (D)---Gene (G)

slide-27
SLIDE 27

27

Experiments: Data

1.

  • D. Davis, N. V. Chawla. Exploring and Exploiting Disease Interactions from Multi-Relational Gene and Phenotype Networks. PLoS One 2011.

2.

  • N. Du, C. Faloutsos, B. Wang, and L. Akoglu. Large human communication networks: patterns and a utility-driven generator. In KDD ’09.

3.

  • Y. Dong, Y. Yang, J. Tang, Y. Yang, and N. V. Chawla. Inferring user demographics and social strategies in mobile social networks. In KDD’14.

k: average degree; cc: clustering coefficient; ac: associative coefficient

Healthcare Networks Disease (D)---Gene (G) Mobile Phone Call Networks Two Operators: Aa---Ab

slide-28
SLIDE 28

28

Experiments: Data

1.

  • D. Davis, N. V. Chawla. Exploring and Exploiting Disease Interactions from Multi-Relational Gene and Phenotype Networks. PLoS One 2011.

2.

  • N. Du, C. Faloutsos, B. Wang, and L. Akoglu. Large human communication networks: patterns and a utility-driven generator. In KDD ’09.

3.

  • Y. Dong, Y. Yang, J. Tang, Y. Yang, and N. V. Chawla. Inferring user demographics and social strategies in mobile social networks. In KDD’14.

k: average degree; cc: clustering coefficient; ac: associative coefficient

Healthcare Networks Disease (D)---Gene (G) Mobile Phone Call Networks Two Operators: Aa---Ab Mobile Phone Call Networks Three Operators: Ea---Eb---Ec

slide-29
SLIDE 29

29

Experiments: Data

k: average degree; cc: clustering coefficient; ac: associative coefficient

Asymmetry Heterogeneity

slide-30
SLIDE 30

30

Experiments: Coupled Networks

1 2 3 4 5 6 7 8 9 10

10 Coupled Prediction Cases

slide-31
SLIDE 31

31

Experiments

AUPR or AUROC or Precision @ k

10 Coupled Prediction Cases

slide-32
SLIDE 32

32

Baselines

1.

  • R. N. Lichtenwalter, J. T. Lussier, and N. V. Chawla. New perspectives and methods in link prediction. In KDD ’10.

2.

  • L. Backstrom and J. Leskovec. Supervised random walks: predicting and recommending links in social networks. In WSDM’11.

AUPR or AUROC or Precision @ k

Unsupervised Methods:

ü Common Neighbors (CN) ü Adamic Adar (AA) ü Jaccard Coefficient (JC) ü Preferential Attachment (PA) ü PropFlow (PF) ü Implicit Target Network (IT)

Supervised Methods:

ü Logistic Regression (LRC)

  • LRC-IT

ü Decision Tree (DT)

  • DT-IT

ü CoupledLP ü CoupledLP-IT

LRC-IT, DT-IT, CoupledLP-IT: NO Implicit Target network construction

features

slide-33
SLIDE 33

33

Experiments

AUPR or AUROC or Precision @ k

Training Links:

ü source links between nodes with cross links ü 1% target links

ü Test Links:

ü 99% target links

slide-34
SLIDE 34

34

Evaluation Metrics

1.

  • R. N. Lichtenwalter, J. T. Lussier, and N. V. Chawla. New perspectives and methods in link prediction. In KDD ’10.

2.

  • L. Backstrom and J. Leskovec. Supervised random walks: predicting and recommending links in social networks. In WSDM’11.

AUPR or AUROC or Precision @ k

ü Area Under Precision Recall Curve (AUPR) ü Area Under Receiver Operating Characteristic Curve (AUROC) ü Precision at Top k

slide-35
SLIDE 35

35

AUPR Results

AUPR

slide-36
SLIDE 36

36

AUROC Results

AUROC

slide-37
SLIDE 37

37

Precision@k Results

Precision @ k

slide-38
SLIDE 38

38

Effects of Implicit Target Network

AUPR

slide-39
SLIDE 39

39

Effects of Implicit Target Network

AUPR

slide-40
SLIDE 40

40

Future Work

  • 1. Efficiency of CoupledLP
  • 2. One-step prediction framework
  • 3. User behavior in coupled networks

… …

slide-41
SLIDE 41

41

Conclusion

Output

Source network Cross network Target network

Input

v Coupled Link Prediction Problem v CoupledLP Framework

Implicit Target Network Construction

  • Solve Incompleteness

Coupled Factor Graph Model

  • Solve Asymmetry
  • Solve Heterogeneity
slide-42
SLIDE 42

42

Thank You!

Questions

Data & Code: https://aminer.org/coupledlp

slide-43
SLIDE 43

43

Effects of Implicit Target Network

x-axis: pruning threshold z y-axis: AUPR / AUROC +5% AUPR +8% AUROC

Unsupervised methods on implicit target network