LarKC WP4 Teleconference, 11.00 -12.20, June 11, 2010
The skype ID of the chair: larkc-wp4
Participants
Reto (STI), Yan (WICI), Vassil (Ontotext), Zhisheng (VUA), Gaston (VUA), Hans (MPG), Lael (MPG), Emanuele (CEFRIEL), Matthias (HLRS), Alex (HLRS), Axel (HLRS), Barry (?OntoText)
Regret: Yi (WICI)
Agenda
- M33 deliverable D4.7.2
- Task 4.4 on Rule-based reasoning (leader: STI Innsbruck)
- * Current status of D4.4.1
- * Ontotext's work on consistency checking by rule-based reasoning (Cooperation with WP7a)
- * Rule-based reasoning for consistency checking on the case study of Korean Traffic sign data (Cooperation with WP6)
- Parallel reasoning (leader: HLRS) and Stream reasoning (leader:CEFRIEL)
- * Parallel reasoning for consistency checking on the case study of Korean Traffic sign data (Cooperation with WP6)
* Parallel reasoning over ?PubMed data and LLD (Cooperation with WP7a)
- Stop rules and reasoning (MPG)
M33 deliverable D4.7.2
D4.7.2 Evolved Evaluation & Revision of plug-ins deployed in use-cases. This deliverable will report on the performance of plug-ins from tasks T4.3 (interleaved reasoning and selection), T4.4 (rule-based reasoning) and T4.6 (stream reasoning), in terms of the targets defined in D4.7.1.
- Guidelines: i) Deploy the plug-ins into the case studies, ii) Use the data sets
from the case studies for the evaluation, iii) use the standard benchmarks (such as Oxford benchmarks and LUBM benchmarks, and iv) cooperate with the SEALS project to make joint effort for the standarization of the evaluation technology.
Reto: The Lehigh University Benchmark is developed to facilitate the evaluation of Semantic Web repositories in a standard and systematic way.
- Task (VUA and MPG): Approximate Reasoning with Heuristics and stop rules
- Task (VUA and WICI): Knowledge summarization technique for interleaving reasoning and selection with PUBMED data and LLD
Task (STI and ?OntoText): Rule-based reasoning and its evaluation
- Task (CEFRIEL): Stream reasoning and its evaluation.
Structure of D4.7.2
- Introduction (ALL)
- Evaluation of Approximate reasoners
- 2.1.Interleaving reasoning and selection with knowledge summarization(VUA and WICI) 2.2 Approximate reasoners with heuristics (VUA and MPG) 2.3 Reasoner with Stop rules (VUA and MPG)
- Evaluation of Rule-based reasoners
- 3.1 STI's work on rule-based reasoners 3.2 Ontotext's work on consistency checking by rule-based reasoning
- Evaluation of Steam reasoning (CEFRIEL)
- Evaluation of Parallel reasoning (HLRS and VUA)
- Conclusion (ALL)
Matthias: Chapter 5 can be optional until we make clear the work become good enough to report as a separate chapter, or report it in Chapter 2, 3, and 4.
Reto: as said, we could do some evaluation about IRIS, which is also done in SOA4All. there will not be any time to do evaluation about the parallelized rule-based reasoner, as the implementaiton for it has not yet started! so section 3.1 might be missing at the end!
Timeline of D4.7.2
June - September: Implementation, Experiments/Data tests, and evaluation
October 1-30,2010, First draft
November 1-30, 2010, Final Integration
December 1-9, 2010, Final Revision
December 10-17, 2010, Deliverable Review
December 18-24, 2010, Quality Control
December 31, 2010, Submit to the Commission.
Task 4.4 on Rule-based reasoning (leader: STI Innsbruck)
- Current status of D4.4.1
- Ontotext's work on consistency checking by rule-based reasoning (Cooperation with WP7a)
Barry: an example:
Consistency: eq-diff1
x <owl:sameAs> y
- [Constraint x != y]
x <owl:differentFrom> y
Another example is the cardinality checking.
- Rule-based reasoning for consistency checking on the case study of Korean Traffic sign data (Cooperation with WP6)
Zhisheng: The cardinality checking would be also useful for the consistency checking for the Korean Traffic Sign Data. They may need the logical consistency checking as well.
Barry: We need more information to see whether or not we can use the same approach of consistency checking on OWLIM for the Korean Traffic Sign Data.
Parallel reasoning (leader: HLRS) and Stream reasoning (leader:CEFRIEL)
- Parallel reasoning for consistency checking on the case study of Korean Traffic sign data (Cooperation with WP6)
Parallel reasoning over ?PubMed data and LLD (Cooperation with WP7a)
Vassil: a typical example of reasoning query over ?PubMed and LLD:
PREFIX skos: <http://www.w3.org/2004/02/skos/core#>
PREFIX lifeskim: <http://linkedlifedata.com/resource/lifeskim/>
PREFIX pubmed: <http://linkedlifedata.com/resource/pubmed/>
SELECT ?title ?respiratoryDisorder ?leukocytes ?obesity
WHERE {
- ?doc pubmed:year '2009' . ?concept1 skos:prefLabel "Leukocytes" . ?concept1a skos:broaderTransitive ?concept1 . ?concept1a skos:prefLabel ?leukocytes. ?concept2 skos:prefLabel "Respiration Disorders" . ?concept2a skos:broaderTransitive ?concept2 . ?concept2a skos:prefLabel ?respiratoryDisorder. ?concept3a skos:prefLabel "Obesity" . ?concept3a skos:prefLabel ?obesity. ?doc lifeskim:mentions ?concept1a . ?doc lifeskim:mentions ?concept2a . ?doc lifeskim:mentions ?concept3a . ?doc pubmed:articleTitle ?title .
}
WICI and VUA are going to test it with the implementation of interleaving reasoning and selection with knowledge summarization.
Stop rules and reasoning (MPG)