Workpackages
- The project has four parallel workpackages:
- TF: TermFactory
- I2I: Information to Intelligence
- MuTOL: Multilingual terminology and ontology learning
- SIS: Service Interoperability Support
- Please click on the workpackage of your choice to see the contents of this package on the task level.
TF: TermFactory
- The TF WP consists of six tasks:
- Representation and combination of multilingual, multi-source terminologies in OWL/RDF.
- Top ontologies and their NLP lexicons.
- Import of vocabularies into TermFactory.
- Recommendations on terminology management interfaces.
- Realisation of interfaces for terminology work.
- Workflows for collaborative distributed terminology work.
- Representation and combination of multilingual, multi-source terminologies in OWL/RDF
- In previous projects (4M, COGKS, TF preparatory project) basic solutions for ontology-based representation of multilingual terminology have been developed. This task completes the solution and checks conformance with relevant terminology and language technology standards.
This task also supplements standards related to sharing ontologies. The support of the ontology language OWL for managing ontology hierarchies is insufficient. There are no primitives for importing/exporting selected concepts from an ontology. - Top ontologies and their NLP lexicons
- This task fleshes out the TF top ontology developed in the preparatory project and the language technology suited i.a. for automatic parsing and generation of NL descriptions of concepts from the ontology.
- Import of vocabularies into TermFactory
-
- Selection or development of standards and conversions for import/export of term records.
- Conversion of a number of vocabularies into TermFactory.
- Recommendations on terminology management interfaces
- Recommendations on the terminology management interface will be provided based on the experience obtained from the development of MyTerMS terminology management programme and analysis of interfaces of other terminology management systems.
- Realisation of interfaces for terminology work
-
- Selection of the platform for carrying out collaborative distributed terminology work.
- Conversion of a number of vocabularies into TermFactory.
- Workflows for collaborative distributed terminology work
- Quality manual and best practice guidelines for terminology work in the TF environment will be provided.
I2I: Information to Intelligence
- The I2I WP consists of five tasks:
- Planning and Specifications
- Initial analysis of the application areas: draw up detailed specifications of deliverables and milestones, in each application area.
- Technology Research
-
- Strengthen the core IE engine and knowledge bases.
- Automatic knowledge acquisition: learning ontologies for the target application domains.
- Automatic knowledge acquisition: learning patterns for the target application domains.
- Information Aggregation.
- Business-Intelligence Application
- Together with our BI partners, we will define approximately 5 sub-tasks/topics to develop over the course of the project. The initial topic will include Investments; further topics may include Advertising campaigns, Acquisitions and Marketing.
- Medical-Intelligence Application
- Medical-intelligence application area involves three major sub-tasks:
-
- Determining epidemic events as accurately as possible.
- Detailed analysis of interference from non-relevant, secondary events.
- Multi-lingual coverage: For the Medical application we will cover one more language beyond English.
- Evaluation Tasks
- Evaluation of Information Aggregation, in the target application areas, involving partners’ participation. Review of the partners’ recommendations.
MuTOL: Multilingual terminology and ontology learning
- The MuTOL WP consists of two tasks:
-
- Evaluate the multilingual keyphrase extraction method Likey with 2-3 different data sets and with different languages. A collection of Wikipedia articles and European parliament proceedings will be used. A method called Morfessor developed earlier in our center will be applied for the segmentation of words.
- Develop an approach for statistical extraction of ontological structures or conceptual hierarchies based on the automatically generated keyphrase collections.
SIS: Service Interoperability Support
- The SIS WP consists of five tasks:
-
- Prepare the backend of the service interoperability support to create logs about the collaboration activities that take place between different parties during ontology creation. These logs must be suitable for applying social mining techniques for mining the collaboration processes.
- Evaluate tools that are candidates for assisting in the creation of ontologies for supporting the TF,. The design tool needs to support both terminologist needs and system designer needs concerning collaboration management and service interfaces. In this area, expertise gained from previous projects can be applied. The evaluation also detects the respective functionality gaps of tools that need to be filled with additional implementation work to render them suitable for adoption into SIS.
- Collect information from the web cloud about services, persons backing services, and organizations that request or provide services.
- Creating and accessing service unification relationships. Evaluate existing web service coordination and reflection mechanisms, and model-driven engineering tools. Collect model-based unification relationships (e.g. model transformations and weaving models) between service implementation and design artifacts and select a set of facilities to feed information about the relationships. Create retrieval methods for application programmers to automatically access that information and utilize them in applications.
- The case study in which organizations use SIS and extensions for ontology building to carry out electronic B2B collaboration.