1 Computing Trust in Social Networks Huy Nguyen Lab seminar April 15, 2011.

18
Computing Trust in Social Networks Huy Nguyen Lab seminar April 15, 2011

Transcript of 1 Computing Trust in Social Networks Huy Nguyen Lab seminar April 15, 2011.

Page 1: 1 Computing Trust in Social Networks Huy Nguyen Lab seminar April 15, 2011.

1

Computing Trust in Social Networks

Huy NguyenLab seminar April 15, 2011

Page 2: 1 Computing Trust in Social Networks Huy Nguyen Lab seminar April 15, 2011.

2

Web-Based Social Networks (WBSNs)

• Websites and interfaces that let people maintain browsable lists of friends

• Last count (2008)– 245 social networking websites– Over 850,000,000 accounts

Page 3: 1 Computing Trust in Social Networks Huy Nguyen Lab seminar April 15, 2011.

3

Using WBSNs

• Lots of users, spending lots of time creating public information about their preferences

• We should be able to use that to build better applications

• When I want a recommendation, who do I ask?– The people I trust

Page 4: 1 Computing Trust in Social Networks Huy Nguyen Lab seminar April 15, 2011.

4

Applications of Trust

• With direct knowledge or a recommendation about how much to trust people, this value can be used as a filter in many applications

• Since social networks are so prominent on the web, it is a public, accessible data source for determining the quality of annotations and information

Page 5: 1 Computing Trust in Social Networks Huy Nguyen Lab seminar April 15, 2011.

5

Research Areas

• Inferring Trust Relationships• Using Trust in Applications

Page 6: 1 Computing Trust in Social Networks Huy Nguyen Lab seminar April 15, 2011.

6

Inferring Trust

The Goal: Select two individuals - the source (node A) and sink (node C) - and recommend to the source how much to trust the sink.

A B CtAB tBC

tAC

Page 7: 1 Computing Trust in Social Networks Huy Nguyen Lab seminar April 15, 2011.

7

Methods

• TidalTrust– Personalized trust inference algorithm

• SUNNY– Bayes Network algorithm that

computes trust inferences and a confidence interval on the inferred value.

Page 8: 1 Computing Trust in Social Networks Huy Nguyen Lab seminar April 15, 2011.

8

SourceSink

Page 9: 1 Computing Trust in Social Networks Huy Nguyen Lab seminar April 15, 2011.

9

TidalTrust Algorithm

• If the source does not know the sink, the source asks all of its friends how much to trust the sink, and computes a trust value by a weighted average

• Neighbors repeat the process if they do not have a direct rating for the sink

Page 10: 1 Computing Trust in Social Networks Huy Nguyen Lab seminar April 15, 2011.

10

SUNNY

• Trust inference algorithm using Bayesian Networks

• Trust network is mapped into a Bayes Net• Conditional probability values are computed

through profile similarity measures• A “most confident” subnetwork is selected

and trust inference is performed on that network

• Result is an inferred trust value and a confidence in that value

Page 11: 1 Computing Trust in Social Networks Huy Nguyen Lab seminar April 15, 2011.

11

Confidence in Social Networks

• P(n|n’): prob that n believes n’• Calculate P(n|n’) based on profile

similarity1.Overall difference Ө2.Difference on extreme χ

3.Maximum difference ∆

4.Correlation coefficient σ

Page 12: 1 Computing Trust in Social Networks Huy Nguyen Lab seminar April 15, 2011.

12

Compute confidence

σ |1 – 2(0.7 Ө + 0.2 ∆ + 0.1 χ ) |

if χ exists

P(n|n’) =

σ |1 – 2(0.8 Ө + 0.2 ∆ ) |

otherwise

Page 13: 1 Computing Trust in Social Networks Huy Nguyen Lab seminar April 15, 2011.

13

Bayesian Network of Trust

• Recursively do– Backward breath-first search from the

source– Forward breath-first search from the

sink

• Final result: set K• Return FAILURE if the source node

is not in K

Page 14: 1 Computing Trust in Social Networks Huy Nguyen Lab seminar April 15, 2011.

14

SourceSink

Page 15: 1 Computing Trust in Social Networks Huy Nguyen Lab seminar April 15, 2011.

15

SUNNY algorithm

1. Build a Bayes Net of the trust domain

2. Compute conditional prob of each node in BN

3. Use the conditional prob to decide if the node is trusted or not

4. Use TidalTrust to compute the trust value

Page 16: 1 Computing Trust in Social Networks Huy Nguyen Lab seminar April 15, 2011.

16

Evaluation: FilmTrust

• Movie recommender• Website has social network where

users rate how much they trust their friends about movies

• Movie recommendations are made using trust

Page 17: 1 Computing Trust in Social Networks Huy Nguyen Lab seminar April 15, 2011.

17

Evaluation

• Movie rating is used to compute confidence values

• SUNNY vs. TidalTrust on FilmTrust network

Page 18: 1 Computing Trust in Social Networks Huy Nguyen Lab seminar April 15, 2011.

18

Conclusions

• Trust is an important relationship in social networks

• Introduced a probabilistic interpretation of confidence in trust network

• Proposed SUNNY an algorithm for computing trust and confidence in social networks