TriQL.P - Trust Architecture

Nearby: W3C Named Graphs Webite | NG4J | TriQL.P QL |

The TriQL.P - Trust Architecture aims at supporting users in their decision whether to trust or to distrust information found on the Semantic Web. The main objective of the architecture is to support a wide range of different context-, content- or reputation-based trust policies. Beside of the actual data, the architecture can also return explanation trees which explain why data should be trusted.

The architecture is based on the Named Graphs data model. The architecture consists of a query language and RDF vocabularies for defining trust policies, for representing explanation trees and for publishing information using digital signatures. It is currently being implemented on top of NG4J.

Note: This page is outdated as the TriQL.P Architecture has been replaced by the WIQA Framework.

1. Introduction

One of the main goals of the Semantic Web initiative is to build an infrastructure which allows applications to use data provided by large networks of independent data sources. Aggregating data from different sources always raises questions about data quality and data trustworthiness. Before the gathered information can be used, its quality and trustworthiness has to be evaluated according to the application-specific trust requirements. This trust decision can be based on:

The main objective of the TriQL.P Trust Architecture is to support a wide range of different, subjective and task-specific trust-policies based on information from all three categories.

The key factor in a user's trust decision is her understanding why given information is fulfilling her trust requirements. Explanation trees try to facilitate this understanding. Explanation trees contain information why query results are matching a given trust policy. Using this information, it is possible to implement something like Tim Berners-Lee's "Oh, yeah?"-button [BernersLee97], meaning that a user can click on every piece of information within an application and get an explanation why she should trust the information.

2. Building Blocks

3. Current Development Status

3.1. Download

Version	Comment	Release Date
V0.01	Initial version submitted to CVS. The engine understands TPL policies and returns query results and explanations for policies that are only using graph patterns. It doesn't support COUNT() and METRIC() yet.	2005-02-16

Version

Comment

Release Date

V0.01

Initial version submitted to CVS.
The engine understands TPL policies and returns query results and explanations for policies that are only using graph patterns. It doesn't support COUNT() and METRIC() yet.

2005-02-16

3.2 Using the TriQL.P Trust Engine

The following code uses a graph set about financial institutions and financial news as example data.

The TPL policy suite below contains several trust policies that applications might want to use in a financial scenario.

The following code loads the example data and policy suite. Afterwards, it queries the data about all rdf:type triples using fin:Policy1 "Trust only information published by rating agencies". The query results are displayed together with their explanations.

....
 
// Load Trust Policy Suite
Model tplFile = ModelFactory.createDefaultModel();
tplFile.read("file:examples/finPolicies.n3", "N3");
PolicySuite suite = new PolicySuiteFromRDFBuilder(tplFile.getGraph()).buildPolicySuite();
 
// Load Example Data
NamedGraphSet source = new NamedGraphSetImpl();
source.read("file:examples/finData2.trig", "TRIG");
 
// Select Trust Policy
TrustPolicy policy = suite.getTrustPolicy("http://www.fu-berlin/suhl/bizer/financialscenario/policies/Policy1");
 
// Run a Query against the Example Data using the Policy
TrustEngine trusted = new TrustEngine(source);
QueryResult result = trusted.find(
  new Triple(Node.ANY, RDF.Nodes.type, Node.ANY),  policy);
 
// Output Query Results
Iterator it = result.tripleIterator();
while (it.hasNext()) {
  System.out.println(it.next());
}
 
// Output Explanations for all Triples
it = result.tripleIterator();
while (it.hasNext()) {
  System.out.println("# ===============================================");
  Triple triple = (Triple) it.next();
  Explanation explanation = result.explain(triple);
  Model m = ModelFactory.createDefaultModel();
  m.getGraph().getBulkUpdateHandler().add(explanation.toRDF());
  m.setNsPrefixes(PrefixMapping.Standard);
  m.setNsPrefix("exp", EXP.getURI());
  m.write(System.out, "N3");
}
 
....

4. Feedback

We are very interested in hearing your opinion about the TriQL.P Trust Architecture. Please send comments to:

5. References

$Date: 2005/01/15 18:34:00 $
$Id: index.html,v 1.1 2005/01/15 18:34:00 bizer Exp $