SWAT abstracts 2004
| SWAT Home |
| People |
| Projects |
| Publications |
| Downloads |
| Contact Info |
In this paper, we present our work on evaluating knowledge base systems with respect to use in large OWL applications. To this end, we have developed the Lehigh University Benchmark (LUBM). The benchmark is intended to evaluate knowledge base systems with respect to extensional queries over a large dataset that commits to a single realistic ontology. LUBM features an OWL ontology modeling university domain, synthetic OWL data generation that can scale to an arbitrary size, fourteen test queries representing a variety of properties, and a set of performance metrics. We describe the components of the benchmark and some rationale for its design.
Based on the benchmark, we have conducted an evaluation of four knowledge base systems (KBS). To our knowledge, no experiment has been done with the scale of data used here. The smallest dataset used consists of 15 OWL files totaling 8MB, while the largest dataset consists of 999 files totaling 583MB. We evaluated two memory-based systems (OWLJessKB and memory-based Sesame) and two systems with persistent storage (database-based Sesame and DLDB-OWL). We show the results of the experiment and discuss the performance of each system. In particular, we have concluded that existing systems need to place a greater emphasis on scalability.
Back to publications page
Y. Guo, Z. Pan, and J. Heflin. An Evaluation of Knowledge Base Systems for Large OWL Datasets. Third International Semantic Web Conference, Hiroshima, Japan, LNCS 3298, Spinger, 2004, pp. 274-288
In this paper, we present an evaluation
of four knowledge base systems (KBS) with respect to use in large OWL
applications. To our knowledge, no experiment has been done with the scale
of data used here. The smallest dataset used consists of 15 OWL files
totaling 8MB, while the largest dataset consists of 999 files totaling
583MB. We evaluated two memory-based systems (OWLJessKB and memory-based
Sesame) and two systems with persistent storage(database-based Sesame and
DLDB-OWL). We describe how we have performed the evaluation and what
factors we have considered in it. We show the results of the experiment
and discuss the performance of each system. In particular, we have
concluded that existing systems need to place a greater emphasis on
scalability.
Back to publications page
J. Heflin, Z. Pan. A Model Theoretic Semantics for Ontology Versioning. Third International Semantic Web Conference, Hiroshima, Japan, LNCS 3298 Springer, 2004, pp. 62-76
We show that the SemanticWeb needs a formal semantics for the various kinds of links between ontologies and other documents.We provide a model theoretic semantics that takes into account ontology extension and ontology versioning. Since the Web is the product of a diverse community, as opposed to a single agent, this semantics accommodates different viewpoints by having different entailment relations for different ontology perspectives. We discuss how this theory can be practically applied to RDF and OWL and provide a theorem that shows how to compute perspective-based entailment using existing logical reasoners. We illustrate these concepts using examples and conclude with a discussion of future work
Back to publications page
Y. Guo and J. Heflin. An Initial Investigation into Querying an Inconsistent and Untrustworthy Web. In Workshop on Trust, Security, and Reputation on the Semantic Web, ISWC 2004.
The Semantic Web is bound to be untrustworthy and inconsistent. In this paper, we present an initial approach for obtaining useful information in such an environment. In particular, we replace the question of whether an assertion is entailed by the entire Semantic Web with two other queries. The first asks if a specific statement is entailed given an identification of the trusted documents. The second asks for the document sets that entail a specific statement. We propose a mechanism for efficiently computing and representing the contexts of the statements and managing inconsistency. This system could be seen as a component in an overall trust system.
Back to publications page
A. Qasem, J. Heflin and H. Muņoz-Avila. Efficient Source Discovery and Service Composition for Ubiquitous Computing Environments. In Workshop on Semantic Web Technology for Mobile and Ubiquitous Applications, ISWC 2004.
To be truly pervasive the devices in a ubiquitous computing environment have to be able to form a "coalition" without human intervention. The Semantic Web provides the infrastructure for discovery and composition of device functionalities. AI planning has been a popular technology for automatic service discovery and composition in the Semantic Web. However, because the Web is so vast and changes so rapidly, a planning agent cannot make a closed-world assumption. This condition makes it difficult for an agent to know when it has gathered all relevant information or when additional searches may be redundant. To avoid redundancy we incorporate Local Closed World reasoning with HTN planning to compose Semantic Web services. In addition, when performing information gathering tasks on the Semantic Web, we use Local Closed World reasoning and a concept of "source relevance" to control the search process. We also describe a prototype agent that we have developed.
Back to publications page
Y. Guo, Z. Pan, and J. Heflin. Choosing the Best Knowledge Base System for Large Semantic Web Applications. Thirteenth International World Wide Web Conference (WWW2004), pp. 302-303, 2004.
We present an evaluation of four knowledge base systems with respect to use in large Semantic Web applications. We discuss the performance of each system. In particular, we show that existing systems need to place a greater emphasis on scalability.
Back to publications page
Z. Pan and J. Heflin. DLDB: Extending Relational Databases to Support Semantic Web Queries. Technical Report LU-CSE-04-006, Dept. of Computer Science and Engineering, Lehigh University, 2004.
We present DLDB, a knowledge base system that extends a relational database management system with additional capabilities for DAML+OIL inference. We discuss a number of database schemas that can be used to store RDF data and discuss the tradeoffs of each. Then we describe how we extend our design to support DAML+OIL entailments. The most significant aspect of our approach is the use of a description logic reasoner to precompute the subsumption hierarchy. We describe a lightweight implementation that makes use of a common RDBMS (MS Access) and the FaCT description logic reasoner. Surprisingly, this simple approach provides good results for extensional queries over a large set of DAML+OIL data that commits to a representative ontology of moderate complexity. As such, we expect such systems to be adequate for personal or small-business usage.
Back to publications page
Heflin, J. and Munoz-Avila, H. Integrating HTN Planning and Semantic Web Ontologies for Efficient Information Integration. Technical Report LU-CSE-04-002, Dept. of Computer Science and Engineering, Lehigh University. 2004.
Back to publications page
J. Heflin, Z. Pan. A Model Theoretic Semantics for Ontology Versioning. Third International Semantic Web Conference, Hiroshima, Japan, LNCS 3298 Springer, 2004, pp. 62-76
We show that the SemanticWeb needs a formal semantics for the various kinds of links between ontologies and other documents.We provide a model theoretic semantics that takes into account ontology extension and ontology versioning. Since the Web is the product of a diverse community, as opposed to a single agent, this semantics accommodates different viewpoints by having different entailment relations for different ontology perspectives. We discuss how this theory can be practically applied to RDF and OWL and provide a theorem that shows how to compute perspective-based entailment using existing logical reasoners. We illustrate these concepts using examples and conclude with a discussion of future work
Back to publications page
Y. Guo and J. Heflin. An Initial Investigation into Querying an Inconsistent and Untrustworthy Web. In Workshop on Trust, Security, and Reputation on the Semantic Web, ISWC 2004.
The Semantic Web is bound to be untrustworthy and inconsistent. In this paper, we present an initial approach for obtaining useful information in such an environment. In particular, we replace the question of whether an assertion is entailed by the entire Semantic Web with two other queries. The first asks if a specific statement is entailed given an identification of the trusted documents. The second asks for the document sets that entail a specific statement. We propose a mechanism for efficiently computing and representing the contexts of the statements and managing inconsistency. This system could be seen as a component in an overall trust system.
Back to publications page
A. Qasem, J. Heflin and H. Muņoz-Avila. Efficient Source Discovery and Service Composition for Ubiquitous Computing Environments. In Workshop on Semantic Web Technology for Mobile and Ubiquitous Applications, ISWC 2004.
To be truly pervasive the devices in a ubiquitous computing environment have to be able to form a "coalition" without human intervention. The Semantic Web provides the infrastructure for discovery and composition of device functionalities. AI planning has been a popular technology for automatic service discovery and composition in the Semantic Web. However, because the Web is so vast and changes so rapidly, a planning agent cannot make a closed-world assumption. This condition makes it difficult for an agent to know when it has gathered all relevant information or when additional searches may be redundant. To avoid redundancy we incorporate Local Closed World reasoning with HTN planning to compose Semantic Web services. In addition, when performing information gathering tasks on the Semantic Web, we use Local Closed World reasoning and a concept of "source relevance" to control the search process. We also describe a prototype agent that we have developed.
Back to publications page
Y. Guo, Z. Pan, and J. Heflin. Choosing the Best Knowledge Base System for Large Semantic Web Applications. Thirteenth International World Wide Web Conference (WWW2004), pp. 302-303, 2004.
We present an evaluation of four knowledge base systems with respect to use in large Semantic Web applications. We discuss the performance of each system. In particular, we show that existing systems need to place a greater emphasis on scalability.
Back to publications page
Z. Pan and J. Heflin. DLDB: Extending Relational Databases to Support Semantic Web Queries. Technical Report LU-CSE-04-006, Dept. of Computer Science and Engineering, Lehigh University, 2004.
We present DLDB, a knowledge base system that extends a relational database management system with additional capabilities for DAML+OIL inference. We discuss a number of database schemas that can be used to store RDF data and discuss the tradeoffs of each. Then we describe how we extend our design to support DAML+OIL entailments. The most significant aspect of our approach is the use of a description logic reasoner to precompute the subsumption hierarchy. We describe a lightweight implementation that makes use of a common RDBMS (MS Access) and the FaCT description logic reasoner. Surprisingly, this simple approach provides good results for extensional queries over a large set of DAML+OIL data that commits to a representative ontology of moderate complexity. As such, we expect such systems to be adequate for personal or small-business usage.
Back to publications page
Heflin, J. and Munoz-Avila, H. Integrating HTN Planning and Semantic Web Ontologies for Efficient Information Integration. Technical Report LU-CSE-04-002, Dept. of Computer Science and Engineering, Lehigh University. 2004.
We integrate HTN planning and Semantic Web ontologies for efficient information integration. HTNs is a hierarchical plan representation that refines high-level tasks into simpler tasks. In the context of information integration, high-level tasks indicate complex queries whereas low-level tasks indicate concrete information-gathering actions such as requests to an information source. Semantic Web ontologies allow software agents to intelligently process and integrate information in distributed and heterogeneous environments such as the world wide web. The integration of HTNs and Semantic Web ontologies allow agents to answer complex queries by processing and integrating information in such environments. We also propose to use local closed world (LCW) information to assist these agents. LCW information can be obtained by accessing sources that are described in a Semantic Web language with LCW extensions, or by executing operators that provide exhaustive information. We demonstrate how the Semantic Web language SHOE can be augmented with the ability to state LCW information.
Back to publications page