Towards a Data-Centric Notion of Trust in the Semantic Web (A Position Statement)

Towards aData-Centric Notion of Trust

in the Semantic Web(A Position Statement)

Olaf HartigDatabase and Information Systems Research Group

Humboldt-Universität zu Berlin

Olaf Hartig - Towards a Data-Centric Notion of Trust in the Semantic Web 2

ExistingResearch

on Trust in the(Semantic) Web


Existing Research on Trust

Trust( )=?

Focus on active entities(e.g. persons, agents, peers)



Trust( )=?

What does it meanto trust an active entity?




● J.B. Barney and M.H. Hansen, 1994: Trust “is the mutual confidence that one's vulnerability will not be exploited.”

Trust( )=?





● L. Mui, M. Mohtashemi and A. Halberstadt, 2002: Trust is the “subjective expectation an agent has about another's future behavior based on the history of their encounters.”

Trust( )=?






Trust ( )=?






● T. Grandison and M. Sloman, 2000: Trust is “the firm belief in the competence of an entity to act dependably, securely, and reliably within a specified context.”

Trust ( )=?






● T. Grandison and M. Sloman, 2000: Trust is “the firm belief in the competence of an entity to act dependably, securely, and reliably within a specified context.”

Trust ( )=?



How do we represent trust?● Binary models (e.g. Golbeck and Hendler, 2004)● Discrete models (e.g. Golbeck et al., 2003)

e.g. highly trusted, moderately trusted, …● Continuous models (Brondsema and Schamp, 2006)

e.g. 0 … 1

Trust ( )=?



How do we calculate trust?● Reputation based models● Network based models (Web of trust)

Trust ( )=?


How do we calculate trust?● Reputation based models● Network based models (Web of trust)


Trust ( )=?

hightrustmoderate

trust

e.g. Guha et al., 2004; Ziegler and Lausen, 2004; Golbeck et al., 2003

?



What is the opposite of trust?

(Recommended: Marsh and Dibben, 2005)

Trust ( )=?

Distrust ( )=?



?

Trust ( )=?

Distrust ( )=? Representations of distrust:

● Part of the trust value (e.g. Brondsema and Schamp, 2006) e.g. -1 … 0 … 1

● Separate value (e.g Victor et al., 2006, Guha et al., 2004)

What is the opposite of trust?

(Recommended: Marsh and Dibben, 2005)


Trust ( )=?


Focus on active entities(e.g. persons, agents, peers)


Shift


... make datathe central subjectof research on trustin the Semantic Web

Let's


… conceiveresearch on trust as an effort

that fits in the area ofinformation quality research

Let's




Let's

Common definition: Fitness for useof information




Let's

● Multidimensional concept




Let's

● Multidimensional concept

Category* Criteria / Dimensions

Intrinsic Accuracy, Objectivity, ...

Contextual Completeness, Relevance, Timeliness, ...

Representational Conciseness, Understandability, ...

Accessibility Availability, Security, ...*Classification by Wang and Strong, 1996


Category* Criteria / Dimensions

Intrinsic Accuracy, Objectivity, ...

Contextual Completeness, Relevance, Timeliness, ...

Representational Conciseness, Understandability, ...

Accessibility Availability, Security, ...*Classification by Wang and Strong, 1996



● Multidimensional concept● Trustworthiness of data is another such IQ criterion

Let's


Trustworthiness of Data

Definition: Subjective belief or disbelief in the truth of the information represented by this data


Trust ( )=?

Focus on Semantic Web Data


Trust ( )=?

Focus on Semantic Web Data

Trust ( )=?

Trust ( )=?

Trust ( )=?http://cia..../Albania

13.2 %

http://.../unemp_rate


Trust ( )=Trust ( )

Source-level Approaches

e.g.: Golbeck and Hendler, 2006; Rowe and Butters, 2009; Zaihrayeu, da Silva, and McGuinness, 2005


Trust ( )=Trust ( )

Source-level Approaches

e.g.: Golbeck and Hendler, 2006; Rowe and Butters, 2009; Zaihrayeu, da Silva, and McGuinness, 2005What if multiple sources provide this data?

What if the data was republished as an aggregation?

What about implicit statements inferred from source data?

Are data sources the only relevant assessment criterion?


Trustworthiness of Data

Definition: Subjective belief or disbelief in the truth of the information represented by this data

● Decision to belief or disbelief affected by:

1. Provenance

2. Information quality

3. Other's opinion


Influence Category: Provenance


Trust ( )=?

Influence Category: Provenance

● How was the creation of the data conducted?● Who or what participated in the creation of the

data and how much do I trust this participant?● To what extend does the input from which

the data was produced represents the truth?● What happened to the data since its creation;

how likely is a manipulation?


Provenance

● provenir (French), meaning to come from

● How a data object in its current state came to be

● This history may start before the creation of the object


Provenance

● Provenance graphs represent the provenance


Provenance

● Basically, everything in a provenance graphhad / might have had an influenceon the data object


Provenance

● Basically, everything in a provenance graphhad / might have had an influenceon the data object

Trust( ) = fct( Trust( ), Trust( ), … )


Influence Category: Information Quality


Influence Category: IQ

● Other IQ criteria may affect our belief in the truth of data i.e. our trustworthiness assessment

● Example: lack of correctness, accuracy, or consistency

➔ Trustworthiness:● Is a more abstract kind of IQ criteria● Comprises multiple other criteria

Trust( ) = fct( Accuracy( ), Currency( ), … )



● Relevancy may depend on context:● Example: Completeness Trust ( )




● Relevancy may depend on context:● Example: Completeness

● Example: Currency

Trust ( )

Trust ( )




● Relevancy may depend on context:● Example: Completeness

● Example: Currency

● Reflect relevancy by weighting the criteria

Trust ( )

Trust ( )

Trust( ) = fct( λAcc

·Accuracy( ), λCur

·Currency( ), … )


Influence Category: Other's Opinion


Influence Category: Opinions

● Similar to recommendation systems● Two options:

● Use opinions of trusted consumers only● Weight opinions by the consumers' trustworthiness

Trust( ) = fct( , , … )


Take-away Summary

● Focus of trust research for the Semantic Web:from an actor-centric view to a data-centric perspective

● Understand trustworthiness of Semantic Web data asan information quality criterion

● Categories of factors that affect the assessment:

1. Provenance

2. Information quality

3. Other's opinion Trust ( )


These slides have been created byOlaf Hartig

http://olafhartig.de

This work is licensed under aCreative Commons Attribution-Share Alike 3.0 License

(http://creativecommons.org/licenses/by-sa/3.0/)

Attribution:● http://www.flickr.com/photos/rrrrred/3809362767/● http://www.hasslefreeclipart.com

Towards a Data-Centric Notion of Trust in the Semantic Web (A Position Statement)

Technology

trust trust

trust iswhat

trust value

opposite of trust

existing research

datacentric notion of

thesemantic web olaf

peers olaf hartig