Information Interpretation and Integration Conference
|
|
- I3CON
The Information Interpretation and Integration Conference (I3CON, prounounced "icon") was the first in what has become a series of workshops that supports the ontology and schema interpretation and integration research communities. The goal of of I3CON was to provide an open forum and a software framework for the systematic, and comprehensive evaluation of ontology/schema interpretation and integration tools. I3CON brought researchers, stakeholders, and technologies together to compare approaches, encouraging the exchange of ideas between industry, academia, and government.
I3CON was modeled after the NIST Text Retrieval Conference (TREC), which has succeeded in driving progress in information retrieval research. Like TREC, I3CON provided test data and a standard format for results. We collated and evaluated the individual results, and held a discussion to assess the results.
I3CON was held at the NIST Performance Metrics for Intelligent Systems (PerMIS) Workshop, on August 25, 2004. The focus of the conference was Automated Ontology Alignment.
I3CON was the result of collaboration between government, industry, and academia, under the coordindation of Lockheed Martin Advanced Technology Laboratories (ATL). The ad hoc Organizational Committee, consisting of highly respected members of academia, industry, and government, contributed vital organizational support.
- I3CON 2004 Schedule
- June 15: Released test ontology pairs to participants
- July 10: Camera-ready papers submitted for PerMIS publication
- July 16: Collected alignment results (in Alignment File format) from participants
- August 25: I3CON special session day at PerMIS
Download the I3CON Experiment Results
View the Presentations
- Ontologies
First, we list all I3CON Experiment Ontologies in the below table as well as post
a zip file with all of them together. Please
let us know if you see any foul namespaces, typos, or unusual code!
We have created equivalent N3 and XML forms of each pair:
you may use either as input to your aligner.
The true alignments are currently being developed and will be released after
submissions for the experiment have been received.
Description |
Ontology A |
Ontology B |
True Alignment |
animalsA.owl from users.ebiquity.org/~hchen4/ont/animals; animalsB.owl is a modified version of animalsA.owl |
animalsA.n3 [owl] |
animalsB.n3 [owl] |
animalsAB.n3 |
soccer.daml is from www.lgi2p.ema.fr/~ranwezs/ontologies and it describes the basic concepts of soccer. basketball.daml is a modified version of soccer.daml which uses basketball concepts rather than soccer concepts |
basketball.n3 [daml] |
soccer.n3 [daml] |
basketball_soccer.n3 |
csA.rdf is from www.cs.umd.edu/projects/plus/DAML/onts/; csB.rdf is from cicho0.tripod.com/. They represent the CS departments at the University of Malta and the University of Maryland respectively |
csA.n3 [rdf] |
csB.n3 [rdf] |
csAB.n3 |
hotelA.owl and hotelB.owl describe the characteristics of hotel rooms. The two ontologies are equivalent, but they describe the same concepts in different ways |
hotelA.n3 [owl] |
hotelB.n3 [owl] |
hotelAB.n3 |
networkA.owl and networkB.owl describe the nodes and connections in a local area network. networkA.owl focuses more on the nodes themselves, while networkB.owl is more encompassing of the connections |
networkA.n3 [owl] |
networkB.n3 [owl] |
networkAB.n3 |
people+petsA.owl from cohse.semanticweb.org/ontologies/people; people+petsB.owl is a modified version of people+petsA.owl |
people+petsA.n3 [owl] |
people+petsB.n3 [owl] |
people+petsAB.n3 |
people+pets-noninstanceA.owl and people+pets-noninstanceB.owl are identical to people+petsA.owl and people+petsB.owl, respectively, except all instance data was deleted. |
people+pets-noninstanceA.n3 [owl] |
people+pets-noninstanceB.n3 [owl] |
people+pets-noninstanceAB.n3 |
russiaA.owl and russiaB.owl describe an overview of Russia including locations, objects, and cultural elements |
russiaA.n3 [rdf] |
russiaB.n3 [rdf] |
russiaAB.n3 |
Second, the table below contains sample ontology pairs with human-posited true alignments.
These are for practice and to illustrate the alignment format we've developed.
We will not be grading your aligners' performance on these ontology pairs.
Description |
Ontology A |
Ontology B |
True Alignment |
Simple wine ontologies derived from Wine.com taxonomy. WineA is in English and WineB is in "babelized" Spanish. |
WineA.n3 |
WineB.n3 |
WineAB.n3 |
Ontologies of military weapons derived from the World Fact Book ontology and the Cyc ontology. |
WeaponsA.n3 |
WeaponsB.n3 |
WeaponsAB.n3 |
More ontology pairs are always welcome.
If you have any that you can contribute to this effort,
please send us an email!
- Participants
The list of participants for the experiment part of I3CON was:
Organization |
Point of Contact |
Research Program |
Lockheed Martin ATL |
Todd Hughes, Ben Ashpole |
Ontology Translation Protocol |
INRIA |
Jerome Pierson |
EXMO |
Teknowledge |
John Li |
Lexicon-based Ontology Mapping |
AT&T |
Lewis Hart |
Intelligent Agent Systems |
University of Karlsruhe |
Marc Ehrig |
Institut AIFB |
- Organizational Committee
The following individuals have provided support for I3CON:
- Todd Hughes, Lockheed Martin Advanced Technnology Laboratories
- Benjamin Ashpole, Lockheed Martin Advanced Techology Laboratories
- David Aha, NRL
- Jerome Euzenat, INRIA
- Tim Finin, University of Maryland Baltimore County
- Nenad Ivezic, NIST
- Deborah McGuinness, Stanford University
- Sergey Melnik, Microsoft
- Adam Pease, Articulate Software
- Larry Reeker, NIST
Acknowledgements
-
The Ontology Translation Protocol (Ontrapro) IR&D Program at Lockheed Martin Advanced Technology Laboratories was the impetus for I3CON. For more information about Ontrapro, contact Todd Hughes.
-
Special thanks to NIST, particularly Larry Reeker and Elena Messina, for hosting I3CON.
|
|
|
|
|
Last Updated 2005-12-30
|
|