Trust Models for Community-aware Identity Management

1
Trust Models for Community-aware Identity
Management

• Choi Hee Chul, Sebastian Ryszard Kruk, Slawomir
  Grzonkowski, Katarzyna Stankiewicz, Brain
  Davis, John G. Breslin

2

• Identity

on the Web
3

• Personal Identity

on the Web
4
Personal Identity

• Account
• Blog
• Email
• Community sites
• Flickr
• Orkut

5

• One URI per one thing

6

• One identity per one person

7
Reason of multiple identities

• Writers want to use a pen name
• People like to grumble on the web
• Makes feel free to write

People want to use
8
Reason of multiple identities

• People want to know one subject not whole
  people

People dont need unique identity
9

• What is the problem?

10
Identity crisis
Sorry, I lost the sources of images
11
Objective

• To solve personal identity crisis

12

• What is the important thingto analyze identity?

13
Link
From http//www.xarxanet.org/xarxanet/novetatsDet
all.xarxanet? xnt_localecaxnt_portal1
nov_novetatId26947
Page Rank
14

• For the personal identity

15
Links
In conflict
Father
Master
16
Trust

• Meta-link to make links
• Can be propagated

Trust
Trust
Trust
17
Trust

• Meta-link to make links
• Can be propagated

Love
Love
Love
Trust
18
Method

• Give a better model of trust
• among the people

19
Trust model of FOAF
Trust Level 1
Trust Level 2
20
Rated relationship model
Trust Value 0.9 0.4 0.36 Trust Level 2
21
FilmTrust (related research)
FilmTrust http//trust.mindswap.org/FilmTrust
22
Complex situation
23
Layered model
24
Person dont need to make multiple identities
25
Layered model
26
Person need to keep one identity
27
Benefits

• Solve identity crisis by unique identity
• Improve propagation of information
• Massage filtering and ordering

28
FOAFRealm

• Single-sign-on system
• Based on FOAF and RDF storage
• P2P/HyperCuP infrastructure
• Semantic Social Collaborative Filtering

29
Bended model
30
Multiple dimension
31
Future work

• Fast calculation algorithm of delegated
  trust without privacy problem
• How to figure out the layers?
• Find meaning of conjunction among the several
  layers

32

• Thank you
• heechul.choi_at_deri.org

33

• Backup slides

34
Capturing trust on the web

• Machine-driven approaches
• PageRank or eBay reputation system
• Calculated as an analogy
• User-driven approaches
• PGP, FOAFRealm
• Needs efforts of user
• Hybrid approaches

35
My identities on web

• Privacy data and life log

• For writing

• For friends

• For research

36
Free to make a new identity
37
Cannot be blocked