J. Hendler, K. Stoffel, and M. Taylor. Advances in high performance knowledge representation. Technical Report CSTR -3672, University of Maryland Institute for Advanced Computer Studies Dept. of Computer Science, 1996.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....sources. The agent uses a set of wrappers and the wrapper induction algorithm STALKER [26] to extract relevant information from the web pages after being shown several marked up examples. When the information is gathered it is stored in the local IEA infobase using Java wrappers on a PARKA [20] knowledgebase. This makes new IEA s fairly easy to create, and forces the difficult parts of this problem back on to KB ontology creation, rather than the production of tools to wrap web pages and dynamically answer queries. Currently, there are some proposals for XML based page annotations ....

....Agent Agent Name Server Interface Agents DomainIndependent Task Agents Task Agents Figure 4: Basic Annotation and Query Agent Organizations themselves. Using a PARKA DB knowledgebase allows efficient, modern relational data storage on the back end and query as well as limited KB inferencing [20]. Task Agen t s . There are two domain task agents; the rest are generic middle agents described earlier. The Annotation Agent directs exactly what information should be annotated for each sequence. It is responsible for storing the raw sequence data, making queries to the various wrapped web ....

J. Hendler and Merwyn Taylor Kilian Stoffel. Advances in high performance knowledge representation. Technical Report CS-TR-3672, University of Maryland Institute for Advanced Computer Studies, 1996. Also cross-referenced as UMIACS-TR-96-56.

....sources. The agent uses a set of wrappers and the wrapper induction algorithm STALKER [18] to extract relevant information from the web pages after being shown several marked up examples. When the information is gathered it is stored in the local IEA infobase using Java wrappers on a PARKA [15] knowledgebase. This makes new IEA s fairly easy to create, and forces the difficult parts of this problem back on to KB ontology creation, rather than the production of tools to wrap web pages and dynamically answer queries. Currently, there are some proposals for XML based page annotations ....

....databases using the existing wrappers and other analysis tools as they are developed, without having to necessarily download and install them themselves. Using a PARKA DB knowledgebase allows efficient, modern relational data storage on the back end and query as well as limited KB inferencing [15]. Task Agents. There are two domain task agents; the rest are generic middle agents described earlier. The Annotation Agent directs exactly what information should be annotated for each sequence. It is responsible for storing the raw sequence data, making queries to the various wrapped web sites, ....

J. Hendler and M. Taylor K. Stoffel. Advances in high performance knowledge representation. Technical Report CS-TR-3672, University of Maryland Institute for Advanced Computer Studies, 1996. Also cross-referenced as UMIACS-TR-96-56.

....sources. The agent uses a set of wrappers and the wrapper induction algorithm STALKER [22] to extract relevant information from the web pages after being shown several marked up examples. When the information is gathered it is stored in the local IEA infobase using Java wrappers on a PARKA [18] knowledgebase. This makes new IEA s fairly easy to create, and forces the difficult parts of this problem back on to KB ontology creation, rather than the production of tools to wrap web pages and dynamically answer queries. Currently, there are some proposals for XML based page annotations ....

....databases using the existing wrappers and other analysis tools as they are developed, without having to necessarily download and install them themselves. Using a PARKA DB knowledgebase allows efficient, modern relational data storage on the back end and query as well as limited KB inferencing [18]. Task Agents. There are two domain task agents; the rest are generic middle agents described earlier. The Annotation Agent directs exactly what information should be annotated for each sequence. It is responsible for storing the raw sequence data, making queries to the various wrapped web sites, ....

J. Hendler and Merwyn Taylor Kilian Stoffel. Advances in high performance knowledge representation. Technical Report CS-TR-3672, University of Maryland Institute for Advanced Computer Studies, 1996. Also cross-referenced as UMIACS-TR-96-56.

....are no guarantees concerning rule correctness or domain completeness. We introduce a model based reasoning approach for solving this problem in which the library designer supplies a domain model of the case library and rules are inferred from the model by the PARKA DB query retrieval system (Hendler et al. 1996). A model is easy to maintain and assures the correctness and completeness of the inferred rules. We have partially implemented this approach in NaCoDAE (Navy Conversational Decision Aids Environment) Breslow Aha, 1997) The following sections describe NaCoDAE, its dialogue inferencing ....

....for their applications, or because maintenance issues complicate the rule updating process. 3. 2 Model Based Dialogue Inferencing Figure 1 summarizes our approach, which integrates NaCoDAE with PARKADB, a high performance knowledge representation system for processing relational queries (Hendler et al. 1996). Instead of constructing a rule set, the library designer interactively enters a library model, composed of an object model (which relates domain objects) and a question model (which relates questions to these objects) For many tasks, these models, represented as semantic networks, will be more ....

[Article contains additional citation context not shown here]

Hendler, J., Stoffel, K., & Taylor, M. (1996). Advances in high performance knowledge representation (Technical Report CS-TR-3672). College Park, MD: University of Maryland, Department of Computer Science.

....into a canonical form. 2. Model Builder: This interactively builds the library model. 3. Rule Generator: This yields text rules, which relate text to hq,ai pairs, and implication rules, which define implications between hq,ai pairs. 4. Parka DB: This is a fast relational querying system (Hendler et al. 1996). It inputs a knowledge base (i.e. set of binary assertions) and a query (i.e. Which new answers can be derived given this problem description ) It outputs a set of newly implied answers to previously unanswered questions. The benefits of this approach derive from the library model s ....

Hendler, J., Stoffel, K., & Taylor, M. (1996). Advances in high performance knowledge representation (Technical Report CS-TR-3672). College Park, MD: University of Maryland, Department of Computer Science.

....making it difficult to create efficient knowledge based systems. For this reason, KDD tools that use ontologies usually pre generalize databases before applying the core data mining algorithm [4, 10, 12] We have developed a tool, ParkaDB that is capable of managing large ontologies. ParkaDB [7, 8] has a very efficient structural design and it is based on high performance computing technologies. Because of its ability to query large ontologies, both serially and in parallel, ParkaDB makes it feasible to merge large ontologies with large databases. The data mining algorithm described in this ....

....ParkaDB s Query Language ParkaDB is knowledge representation system developed by the PLUS group at the University of Maryland. It supports the storage, loading, querying and updating of very large ontologies, both traditional, and hybrid. In this section we describe ParakaDB s query language. See [7, 8] for a discussion on ParkaDB in general. ParkaDB supports a conjunctive query language in which every conjunct is a triple (P; D;R) P can be one of several predefined structural relationships, isa , instanceOf , subCat , etc. Alternatively, P can be an attribute defined on a database. D and R ....

James Hendler, Kilian Stoffel, and Merwyn Taylor. Advances in high performance knowledge representation. Technical Report CS-TR-3672, University of Maryland @ College Park, August 1996.

....new SQL3 standard includes recursion tems provide the generalization operation required by MLClass they have been plaque by their inability to manage large knowledge bases, a requirement for data mining in real world settings. Recently, researchers have initiated projects that address this issue [18, 19, 20, 21]. Generally, the strategy has been to use a DBMS to store data that a KR could retrieve on demand. By using a DMBS to store run time data, KR systems should be able to manage knowledge bases that are as large as the sizes of DB s that a DBMS s can manage, usually available disk space. To mine ....

....by a DBMS should provide the foundation required by MLClass. ParkaDB, created by the PLUS group at the University of Maryland, is the underlying data management system used by MLClass. It is a frame based KR system that uses a generic DBMS to store data on disk instead of in memory at run time [21]. ParkaDB provides the generalization processing that is essential to MLClass. ParkaDB s structure matcher can evaluate the complex queries generated by MLClass. It also provides informative feedback information that MLClass can use to refine queries that ParkaDB does not find answers to. MLClass ....

James Hendler, Kilian Stoffel, and Merwyn Taylor. Advances in high performance knowledge representation. Technical Report CS-TR-3672, University of Maryland @ College Park, August 1996.

No context found.

J. Hendler, K. Stoffel, and M. Taylor. Advances in high performance knowledge representation. Technical Report CSTR -3672, University of Maryland Institute for Advanced Computer Studies Dept. of Computer Science, 1996.

No context found.

Hendler, J., Sto#el, K., , Taylor, M.: Advances in high performance knowledge representation (1996)

No context found.

J. Hendler, K. Stoffel, and M. Taylor. Advances in high performance knowledge representation. Technical Report CSTR -3672, University of Maryland Institute for Advanced Computer Studies Dept. of Computer Science, 1996.

No context found.

Hendler, J., Stoffel, K., Taylor, M.: Advances in high performance knowledge representation. Technical Report CS-TR-3672, University of Maryland Institute for Advanced Computer Studies Dept. of Computer Science (1996)

No context found.

J. Hendler, K. Stoffel, and M. Taylor. Advances in high performance knowledge representation. Technical Report CSTR -3672, University of Maryland Institute for Advanced Computer Studies Dept. of Computer Science, 1996.

No context found.

Hendler, J, Stoffel, K., and Taylor, M. Advances in High Performance Knowledge Representation. University of Maryland Institute for Advanced Computer Studies Dept. of Computer Science, Univ. of Maryland, July 1996. CS-TR-3672 (Also cross-referenced as UMIACS-TR-96-56) 10

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC