trust and semantic web technologies

18
Trust and Semantic Web Technologies Chris McConnell April 4, 2006

Upload: lesley-levy

Post on 31-Dec-2015

26 views

Category:

Documents


1 download

DESCRIPTION

Trust and Semantic Web Technologies. Chris McConnell April 4, 2006. Two Ways to think about Trust. Trust in terms of Web Services Trust (or reputation) on the read/write Web. Trust and Web Services. Trust sits atop Web Services stack Web Services technologies - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Trust and  Semantic Web Technologies

Trust and Semantic Web Technologies

Chris McConnell

April 4, 2006

Page 2: Trust and  Semantic Web Technologies

Two Ways to think about Trust

• Trust in terms of Web Services

• Trust (or reputation) on the read/write Web.

Page 3: Trust and  Semantic Web Technologies

Trust and Web Services

• Trust sits atop Web Services stack

• Web Services technologies

• Needed to protect against malicious users, fraud, flaky business partners.

• Currently, no standard exists for trust in Web Services, so research is speculative.

Page 4: Trust and  Semantic Web Technologies

How could trust be implemented?

• Most articles suggest PKI implementations for authentication.

• As Daconta suggests, current authentication strategies are designed for 1:1 relationships.

• Web Services rely on complex relationships between services (UDDI, WSDL, APIs) more complex than 1:1

Page 5: Trust and  Semantic Web Technologies

Additional Barriers to Trust

• As several authors point out, keys are keys and not users.

• Technological solutions do not guarantee that users are who they say they are.

• Perhaps the most difficult social construct to implement in software.

Page 6: Trust and  Semantic Web Technologies

An RDF approach to Trust

• Uses FOAF and PKI to establish relationships for trusted interactions online.

• Used to sign RDF documents and establish

• Uses a third party to authenticate keys.

Page 7: Trust and  Semantic Web Technologies

Trust and Web 2.0

• Issues of trust in a more explicitly social sphere.

• The “read/write Web” requires trust - or at least reputation” in order to maintain integrity of information or discussion.

• In these cases, it’s not a matter of keeping things private, but instead getting assurance about the quality of public information.

Page 8: Trust and  Semantic Web Technologies

Reputation

• Reputation is based on feedback from other users.

• In offline world, reputation is generally informal

• Online, reputation can be informal or formal.

Page 9: Trust and  Semantic Web Technologies

A formal “Web 1.5” reputation system

• Slashdot uses a “karma” system to rate the reputation of users.

• When users leave comments on entries, these comments can be numerically rated by moderators.

• The sum of these moderation scores determines “karma.”

• Users must reach a particular karma threshold before they can get moderation privileges.

Page 10: Trust and  Semantic Web Technologies

Why was this system developed?

• Slashdot discussions rapidly grew out-of-hand, filled with junk posts, spam, and flamebait.

• Comment ratings allow readers to filter out only the best comments.

• Moderation privileges first went to users known by administrators, then randomly chosen users, until finally settling on current karma system.

Page 11: Trust and  Semantic Web Technologies

Problems with this System

• Initially karma was represented as a numerical value.

• Some users became obsessed with karma: “karma whores”

• New commenters are often ignored, alienated in the moderation system.

• Replicates existing Slashdot attitudes, a self-reinforcing system.

Page 12: Trust and  Semantic Web Technologies

Reputation on Wikipedia• Wikipedia does not have a formal

reputation system like Slashdot.

• Leaders of the project want to encourage as much participation as possible.

• Relies on informal reputation. Contributions to individual articles, participation in Wikiproject, talk pages.

• Vandals can be banned, have their user accounts frozen

Page 13: Trust and  Semantic Web Technologies

Issues for Wikipedia

• Information quality: How can we know this is good information if we don’t know the users?

• Allows anonymous edits, can encourage vandals.

• “Given enough eyeball…”• Reputation is an ancillary issue if many

people are checking pages.

Page 14: Trust and  Semantic Web Technologies

Seigenthaler Incident

• Article on journalist John Seigenthaler accused him of participating in the JFK assassination.

• Posted by an anonymous user.• Article went unnoticed until Seigenthaler

publicized the story in the mainstream media.• Wikipedia response: barring anonymous

users from creating new articles.

Page 15: Trust and  Semantic Web Technologies

Other Ongoing Issues

• Political staffers editing the boss’ article to remove unflattering information.

• Adam Curry editing “Podcast” article to make it more favorable to him.

• Articles that receive little attention can have errors that go unnoticed for long periods.

Page 16: Trust and  Semantic Web Technologies

Future of Trust on Wikipedia

• Jimmy Wales has said publicly that he does not believe the project needs a Slashdot-style

• To improve trust, he says review processes will be expanded.

• Create “gold” and “dev” versions of Wikipedia.

Page 17: Trust and  Semantic Web Technologies

Other Issues of Trust on Web 2.0

• del.icio.us: what happens when spam hits a critical mass on social bookmarking systems?

• Astroturf/FUD blogs. How can blogs be trusted beyond informal social reputation?

• Gaming Digg, etc.

Page 18: Trust and  Semantic Web Technologies

Discussion