assessment of multi-hop interpersonal trust in social networks … · 2015-03-27 · assessment of...

Post on 23-Jul-2020

3 Views

Category:

Documents

0 Downloads

Preview:

Click to see full reader

TRANSCRIPT

Assessment of Multi-Hop Interpersonal

Trust in Social Networks by 3VSL

Guangchi Liu, Qing Yang, Honggang Wang, Xiaodong Lin and Mike P. Wittie

Presented by Guangchi Liu

Department of Computer Science

Montana State University

Bozeman, MT, USA

E-mail: guangchi.liu@msu.montana.edu

IEEE INFOCOM 2014, Toronto, CA

2

Motivation

Ask Yourself:

How to trust someone you have never known before?

3

?

Alice David

Trustor Trustee

Motivation

Probably, If…

4

Alice David

Trustor Trustee

Bob

Motivation

“Transitivity of Trust[1]”

[1] Falcone, Rino, and Cristiano Castelfranchi. "Transitivity in trust a discussed property." (2010).

Motivation

Then, How Should Alice Trust David ?

5

Alice David

Trustor Trustee

Bob

?

Complex Topology: Harder Cases

6

?

Parallel

Topology

Motivation

?

Bridge

Topology

Arbitrary

Topology

?

7

Motivation

Finally, we could expect….

Small World Network

Six degrees of separation

8

Applications

9

Consumer-to-Consumer Recommendation Systems

Untrustworthy User

Trustworthy User

Buyer

Seller

Applications

10

Active Friending in OSNs[2]

Trustworthy User

Finder

Target User

Strong Relation

Weak Relation

[2] Yang, De-Nian, et al. "Maximizing acceptance probability for active friending in online social networks."

Proceedings of the 19th ACM SIGKDD international conference on Knowledge discovery and data mining. ACM, 2013.

Applications

11

Applications

Sybil Identification

Untrustworthy User

Trustworthy User

User

Unknown user

?

12

Contributions

13

Contributions

Methodology Three-Valued Subjective Logic (3VSL) Multi-hop trustworthiness assessment

Advantage Applicable to Complex Topology Be able to handle any large scale social network

Validation Real World Experiment 100+ participants in online survey

14

Related Works

15

Related Works

3VSL

Bayesian Network

Hidden Markov Model

Subjective Logic

Graph Partitioning

Random Walk

Community Analysis

Artificial Intelligence

Methods

Graph Theory

Methods

16

Three-Valued Subjective Logic (3VSL)

17

3VSL: Opinion Vector

Positive Negative Uncertain Impression

Interpersonal Trust: Is David telling the truth?

Opinion

David

Trustee Trustor

Alice

18

3VSL: Opinion Vector

Mom Alice’s Evidences: Posteriori Evidence

Positive Negative Uncertain Impression

? Interaction History

19

3VSL: Opinion Vector

Stranger Alice’s Evidences: Priori Evidence

Positive Negative Uncertain Impression

Cognition in Mind

20

3VSL: Opinion Vector

Positive Negative Uncertain Impression

+ + +

Priori Evidence

Posteriori Evidence

21

Dirichlet Distribution:

1 2

1 1 1

1 2 1 2

( , , , )

1(1 )

( , , )

f p p

p p p pB

Positive

Negative

Uncertain

1p

2p

3p

#

#

#

3 1 2: 1P P P

+

3VSL: Opinion Vector

22

Positive

Negative

Uncertain

1

2

3

( )

( ) =

( )

E P

E P

E P

# # #+ + #

# # #+ + #

# # #+ + #

1 1 1

1 2 1 2 1 2

1( , , , ) (1 )

( , , )f p p p p p p

B

3VSL: Opinion Vector

23

1( )E P

Positive Negative Uncertain Impression

2( )E P 3( )E P

# # #+ +

Positive Negative Pos-uncertain

# # #+ + # # #+ + # # #+ +

+ +

# # #+ + #

# # #+ + #

# # #+ + #

3VSL: Opinion Vector

24

Opinion Vector

[ , , , , ]A A A A A A

X X X X X Xb d n e a

A X

Opinion

A

XbA

XdA

Xn A

XaImpression

A

Xe

Positive Negative Pos-uncertain

3VSL: Opinion Vector

“1”

25

Expected Belief

1( )

3

A A A A A

X X X X XE b a n e

Impression Positive Negative Pos-uncertain

A X

3VSL: Opinion Vector

26

Is David telling the truth?

Opinion of David’s Mom

Opinion of a Stranger

0.442 0.45 0.5 0.05

0.1 0.5

( ) 0.33E

( ) 0.33E

3VSL: Opinion Vector

27

G(V, E, W)

v u ?

G(V, E, W)

vi vj

e

w

Multi-hop Trust Computation in OSNs:

28

3VSL: Discounting Operation

Operation for Trust Propagation

29

3VSL: Discounting Operation

Alice David Bob

Positive

Negative

Uncertain

Positive

Negative

Uncertain

A

BB

D

A

D

30

Negative

?

?

A

DA

BB

D

Positive

Uncertain

Negative

3VSL: Discounting Operation

31

Alice David Bob

A

BB

D

A

D

Distortion

B

D

3VSL: Discounting Operation

32

Alice

David Bob

A

D

B

D

David

2 Hops

… Hop #

Absorbing Process

1 Hop

3VSL: Discounting Operation

33

( , ) 1

A A B

D B D

A A B

D B D

A B A A A B

B D D D D D

A B

D D

A A

D D

b b b

d b d

n b d e

e e

a a

A

BB

D

( , )A A B

D B D

Alice Bob David

[ , , , , ]B B B B B B

D D D D D Db d n e a

[ , , , , ]A A A A A A

B B B B B Bb d n e a

Discounting Operation: ( , )A B

B D

3VSL: Discounting Operation

34

Operation for Trust Fusion

3VSL: Combining Operation

35

Alice David

2A

D

1A

D

A

D

Positive

Negative Uncertain

Positive

Negative Uncertain 1A

D

2A

D

3VSL: Combining Operation

36

Positive Pos-uncertain 1A

D

1A

D

+ +

Negative

+

Positive Pos-uncertain Negative

Independent

Evidences

Dependent

Evidences

Positive Negative Pos-uncertain

Posteriori Evidences

Positive Negative Pos-uncertain

Normalize

Priori Evidences

A

D

3VSL: Combining Operation

37

1 2 2 1

1 2 2 1

1 2 2 1

1 2 2 1

1 2 2 1

1 2

1 2 2 1

2 1

1 2 2 1

( , )

A A A AA D D D DD A A A A

D D D D

A A A AA D D D DD A A A A

D D D D

A A A AA A A D D D DD D D A A A A

D D D D

A AA D DD A A A A

D D D D

A A

D D

b e b eb

e e e e

d e d ed

e e e e

n e n en

e e e e

e ee

e e e e

a a

1 2( , )A AA

D D D

Alice David

2 2 2 2 2 2[ , , , , ]A A A A A A

D D D D D Db d n e a

1 1 1 1 1 1[ , , , , ]A A A A A A

D D D D D Db d n e a

Combining Operation: 1 2( , )A A

D D

1A

D

2A

D

3VSL: Combining Operation

38

Method 1:

Step.3

Step.1

Step.2

Step.1

Step.2

Method 2:

Equivalent

1 2( , ( , )B BA

B C C 1 2( ( , ), ( , ))B BA A

B C B C

3VSL: Property

39

3VSL: Topologies

Bridge Topology Arbitrary Topology

Large Social Networks

40

Evaluation

41

Evaluation: Real World Validation

Accuracy of 3VSL in Real Social Network

) – (

1. 2.

1.

2.

: Online Questionnaire[3,4] (100+ samples)

Goal

Method

Topology

[3]. Xiaoming Li and Guangchi Liu. Online Trust Survey. http://www.cs.montana.edu/yang/trust-us, 2013.

[4]. C. Johnson-George and W. C. Swap, “Measurement of specific interpersonal trust: Construction and validation of a scale to assess trust in a specific other.”

Journal of Personality and Social Psychology, vol. 43, no. 6, pp. 1306–1317, 1982.

( ) Error =

) – ( ) ( Error =

42

Evaluation: Real World Validation

43

Evaluation: Numerical Analysis

3VSL Vs. Common Sense

1. 2.

Goal

Method

Topology

3.

? ( )E

44

Evaluation: Numerical Analysis

? ( )E

45

Evaluation: Numerical Analysis

? ( )E

2 1

46

Evaluation: Numerical Analysis

? ( )E

47

Conclusion

48

Conclusion

Three-Valued Subjective Logic

Thank You !

Q & A

IEEE INFOCOM 2014, Toronto, CA

Guangchi Liu (Luke)

刘光迟 (劉光遲)

Networking Lab, Dept. of Computer Science

Montana State University, MT, United States

LinkedIn: www.linkedin.com/pub/guangchi-liu/60/a96/961/

E-mail: guangchi.liu@msu.montana.edu

50

Backup I

Trust

Trust

(Cognitive)

Trust

(Probabilistic)

51

Backup II

Trust Assessment

(AI Methods)

Subjective Logic

(SL)

Bayesian Network

HMM Model

Multi Hop Trust Computation Jøsang, Audun.

"A logic for uncertain probabilities."

Single hop trust assessment Wang, Yao, and Julita Vassileva.

"Trust and reputation model in peer-to-peer networks."

Dynamic of Trust ElSalamouny, Ehab, Vladimiro Sassone, and Mogens Nielsen.

"HMM-based trust model."

52

Backup III

Subjective Logic 3VSL

1 2( , )

Collapse

1 2( , )

Constant

Large Social

Networks

Large Social

Networks

top related