28
Reputation & Trust The copyright of images belong to their authors. I will remove them on demand. Contact me at adrian.holzer@epfl.ch Denis Gillet Adrian Holzer

Trust & reputation

Embed Size (px)

DESCRIPTION

 

Citation preview

Page 1: Trust & reputation

Reputation & Trust

The copyright of images belong to their authors. I will remove them on demand. Contact me at [email protected]

Denis GilletAdrian Holzer

Page 2: Trust & reputation

“what is generally said or believed about a person’s or thing’s character.”

Reputation

Page 3: Trust & reputation

“the extent to which one is willing to depend on another in a situation with a feeling of security, even in the face of possible negative consequences.”

TRUST

Page 4: Trust & reputation

How is trustworthiness measured?

Page 5: Trust & reputation

Trust in physical world

Derived from

personal experience

Influenced by received referrals from others

Trust related information

shared within local

communities

Page 6: Trust & reputation

Trust in online environments Real-life evidences of trust are missing

Need adequate

electronic substitutes

for traditional cues

Trust related information shared on a global scale

Page 7: Trust & reputation

A reputation score is associated with an item

The score is calculated by aggregation of all people’s trust opinions

The score is visible to the entire community

The score represents the trustworthiness

Page 8: Trust & reputation

Examples

Page 9: Trust & reputation

The reputation system incites sellers to be honest

Buyers use sellers’ reputation scores to assess the quality of sellers’ service

Sellers gain reputation when receiving positive ratings

Buyers provide a rating score to sellers

Page 10: Trust & reputation
Page 11: Trust & reputation
Page 12: Trust & reputation
Page 13: Trust & reputation
Page 14: Trust & reputation
Page 15: Trust & reputation

Rank reviewers (Advisor, Top Reviewer, Category Lead) Trust network (trust/distrust a person)

Page 16: Trust & reputation

a professional community for eGovernment, eInclusion, eHealth

Page 17: Trust & reputation
Page 18: Trust & reputation
Page 19: Trust & reputation

(L. Page, “The PageRank Citation Ranking: Bringing Order to the Web”, Technical Report, 1998)

Page 20: Trust & reputation

Define different contexts

Specify a user’s reputation score depending on a

particular context

Page 21: Trust & reputation

A user’s reputation score in Computer Science is high

Her reputation score in Chemistry is low

Page 22: Trust & reputation

with current systems

Problems

Page 23: Trust & reputation

Raters don’t directly benefit from rating items

Low incentive for providing rating

Page 24: Trust & reputation

!

People hope to get a positive rating in return

Positive rating bias

Page 25: Trust & reputation

!

Some people provide unfair ratings due to personal reasons

Unfair ratings

Page 26: Trust & reputation

People with low reputation scores change their identities and enter the community as newcomers

Change of identities

Page 27: Trust & reputation

POSSIBLE SOLUTIONS

Page 28: Trust & reputation

Bias toward positive ratingsAllow anonymous ratings

Give extremely low reputation score to newcomers

Effective incentive mechanisms are needed

Exclude or give low weight to presumed unfair ratings, using statistical analysis

Low incentive for providing rating

Unfair ratings

Change of identities