Disclaimer: This is the local web site for
the PARMENIDES Project at the University of Zurich, for internal use.
Disclaimer: This is the local web site for
the PARMENIDES Project at the University of Zurich, for internal use.
Parmenides believed everything must exist, which meant to
him that change was an optical illusion of some kind. Since both past and
future already exist, he argued that the passing of time must be unreal.
And so Parmenides denied change, saying it was appearance only, and taught
that existence or being is one.
Existence could not be created and was indestructible.
He may have been the first western philosopher to describe the universe as
a permanent single whole, rather than a product of many parts.
From: Parmenides
and Zeno
Ontology driven Temporal Text mining on organisational data for extracting temporal valid knowledge.
Proposal No: IST-2001-39023
Call Identifier: IST-02-8-1A
Research Programme(s): IST-2001
Contact Person: Dr. Babis Theodoulidis
Telephone No:(004-4161)2003309
Fax No: (004-4161)2003324
E-mail: babis@co.umist.ac.uk
Abstract:
Parmenides is focussed on organisational knowledge management , as the
primary aim is to develop an ontology driven systematic approach for integrating
the entire process of information gathering, processing and analysis. Although
project is speculative, it is based on established tools derived from applied
research in temporal text mining, information extraction and ontologies,
to reduce the overall project risk. Parmenides will design and develop a number
of components that together will enhance in a novel way a user s ability
to extract temporal valid knowledge, mostly automatically, from web-based
or conventional textual sources. The process is driven by ontologies that
are built semi-automatically.
Duration (in Months): 30
Keywords: Text Mining, Ontology, NLP
Objectives:
The Parmenides project will develop novel techniques for the purpose
of (a) Semi-automatically building and maintaining domain specific ontologies.
(b)Automatic detection and extraction of events in textual web data. (c)Integrating
the textual temporal information which has been extracted, in a document
warehouse. (d) Knowledge discovery tasks for trends analysis, temporal validity
of knowledge etc. Three major case studies will be used to evaluate and demonstrate
the framework. One for Unilever Nederland BV, the other for Biovista and the
third for MOD. An additional objective is to test and refine ISO/EAGLES evaluation
guidelines by applying these in a realistic environment.
Description of the work:
The Parmenides consortium consists of three IT universities, one research
institution, one business university, two software houses, one multinational
manufacturer, a knowledge/information product company and one public organisation(knowledge
management Department). It will run for 30 months. The work is organised
into nine workpackages. WP1 is concerned with project management.WP2 is concerned
with detailed requirements analysis.WP3 defines the overall component-based
architecture, of the Parmenides project. The result of this work package
will be a detailed description of the components of the system and the interfaces
between them. WP4 is concerned with the development of ontology semi-automatically
in order to support the case studies.WP5 deals with the identification and
extraction of conceptual information from documents, through the application
of information extraction and lexical chaining techniques to individual documents.WP6
deals with the implementation of the Document Warehouse. WP7 is where the
implementation of temporal knowledge discovery component takes place including
the algorithms for analysing text and data, the interfaces to the document
warehouse and finally, the end user interfaces. WP8 is concerned with the
case study reports explaining the need and the requirements. The work also
addresses the definition of the evaluation framework in terms of criteria
and finally, the actual results of the evaluation based on the initial requirements
of the two users of the consortium. WP9 is concerned with consolidating
the lessons of the project in the methodology of evolving the knowledge
management capabilities of the organisation, and with disseminating and
exploiting the project s outcomes.
Milestones and expected results:
M1.The case study requirements, evaluation framework and infrastructure
are in place, and the document collector is implemented. M2.The document
warehouse is implemented and the domain ontology is substantially complete.
Following that integration with the other tools can be tested and finalised.M3.The
evaluation of the tools in the case study context has been conducted, and
the final, enhanced user tools are completed. The exploitation plans are also
finalized.
Partners:
CO0 UMIST (Manchester,
UK), Prime Contractor
CR2 Wordmap Ltd. (Bath, UK)
CR3 Neurosoft S.A. (Athens, EL)
CR4 Otto-von-Guericke Universität Magdeburg
(Magdeburg, DE)
CR5 Unilever Nederland BV (Vlaardingen,
NL)
CR6 Biovista (Athens, EL)
CR7 ISSCO (Genève, CH)
CR8 Zurich University (Zurich,
CH)
CR9 MOD (Athens, EL)
Plus, as subcontractors:
CNRS (Paris, F)
University of Salford
(Manchester, UK)
James Dowdall, Jeremy Elleman, Michael Hess, Will Lowe,Fabio Rinaldi. The role of MultiWord Terminology in Knowledge Management. LREC-2004, Lisbon, 24th-30th May 2004.
Kaarel Kaljurand, Fabio Rinaldi, James Dowdall, Michael Hess. Exploiting Language Resources for Semantic Web Annotations. LREC-2004, Lisbon, 24th-30th May 2004.
Fabio Rinaldi, James Dowdall, Michael Hess, Jeremy Ellman, Gian Piero Zarri, Andreas Persidis, Luc Bernard, Haralampos Karanikas. Multilayer annotations in Parmenides. The K-CAP2003 workshop on "Knowledge Markup and Semantic Annotation", October 25-26, 2003. Sanibel, Florida, USA.
Fabio Rinaldi, Kaarel Kaljurand, James Dowdall, Michael Hess. Breaking the Deadlock. ODBASE 2003 (International Conference on Ontologies, Databases and Applications of SEmantics) Catania, Italy. [LNCS 2888, pp. 876-888., Springer Verlag, 2003]
Fabio Rinaldi, James Dowdall, Michael Hess, Kaarel Kaljurand, Andreas Persidis, Babis Theodoulidis, Bill Black, John McNaught, Haralampos Karanikas, Argyris Vasilakopoulos, Kelly Zervanou, Luc Bernard, Gian Piero Zarri, Hilbert Bruins Slot, Chris van der Touw, Margaret Daniel-King, Nancy Underwood, Agnes Lisowska, Lonneke van der Plas, Veronique Sauron, Myra Spiliopoulou, Marko Brunzel, Jeremy Ellman, Giorgos Orphanos, Thomas Mavroudakis, Spiros Taraviras. Parmenides: an opportunity for ISO TC37 SC4? The ACL-2003 workshop on Linguistic Annotation, July 2003, Sapporo, Japan.
(April 2003)
Related Annotations Schemes
TERQAS
(Proposed Standard for Temporal Annotation)
MUC7 Named Entities Task
MUC7 Co-Reference Task
MUC7 IE Task
MUC-7 Proceedings
GDA initiative
Text Encoding Initiative
GENIA project
MATE
(co-reference in dialogue)
GNOME (NPs
for Language Generation)
EAGLES
ISLE
Corpus Encoding Standard
OLIF
MULTEXT
Ontologies
DAML
Basic references
The XML Specification
Resource Description Framework
Other
The First NLP and XML
Workshop
The
Second NLP and XML Workshop
The Third
NLP and XML Workshop
Nancy Ide
Links to internal resources useful for the project.
Zurich presentations at the Vlaardinger Meeting (Nov 02): (powerpoint, tar-zipped) Our report on the potential role of TimeML in Parmenides: (postscript file)