Workpackage description

Workpackage 1
Workpackage 2
Workpackage 3
Workpackage 4
Workpackage 5
Workpackage 6
Workpackage 7
Workpackage 8
Workpackage 9
Workpackage 10
Workpackage 11

WP-1: Developing tools for Ontology Construction and Interchange
Go to the list
Ontologies guide each of the various means of On-To-Knowledge. Each layer of its tool environment, the methodology, and each of these case studies is organised around the use of ontologies. Therefore, advanced support in developing ontologies is a key asset in guarantying the success of On-To-Knowledge. This support has to be provided in three different ways: developing ontologies from scratch, adapting (i.e., reuse) existing ontologies for new applications, and getting ontologies for free, i.e., applying automatic ontology construction techniques.
WP-1.1 Ontology Language
In order to work with ontologies they must be expressed by using some formalism with predefined epistemological primitives. We will develop a language for describing ontologies using and adapting existing proposals.
WP-1.2 Ontology Editor
An advanced visualization editor will be developed that will ensure that ontology can be performed by information providers who are not IT experts.
WP-1.3 Evolving Ontologies
Ontologies need to change over time. To manage this process, we need to explicitly model their relationships with earlier ontology versions. Maintaining the change solves the legacy problem when older versions of ontologies where used to annotate large bodies of information. On-To-Knowledge provides version management for ontologies integrated with automatic translation between the different ontologies. These translations are defined with the ontology translator.
Existing ontologies change when they are adapted to the task- and application-specific needs. We will provide a library of reusable ontologies with support for selecting and adapting these ontologies
WP-2: Developing support in Information extraction
Go to the list
This Workpackage deals with the automatic generation and refinement of ontologies as well as parts of them. Autonomous ontology creation is of central interest to any project that deals with ontology-based information management.
WP-2.1 Exploration and Evaluation of existing information extraction techniques
Exploration and Evaluation of already known ways for performing the task of information extraction from various information sources (this will possibly mean discussion and interaction with other European project groups such as the TREVI project).
WP-2.2 Information extraction techniques for unstructured documents
Refinement and adaptation of the current approach to analysis and retrieval of semantic knowledge from unstructured documents w.r.t. the ontology approach undertaken in the On-To-Knowledge project and the results of WP-2.1 defined above.
Ontology based concept extraction: a natural language based tool for extraction ontological concepts from documents. This workpackage will exploit much of the current in-house technology of CognIT.

Capturing of "general knowledge" as well as "specific knowledge". The roles of "names" in a text structure has to be clarified.
WP-2.3 Information extraction techniques for semi-structured documents (wrappers)
Ontology supported wrappers: a pattern-matching language for adding structure to documents based on patterns found in the documents and on what is required for ontological categories. Existing technology that will be exploited here are approaches such as W4F, JEDI, and in-house technology of BT.
  • Wrapper definition tool. A tool for writing such wrappers.
  • Wrapper application tool. An engine for applying such wrappers to documents; either off-line or on line (i.e., at query-time). This engine should also include annotations stemming from WP2.2.
WP-3: Developing an infrastructure for Information representation
Go to the list
The purpose of the Representation Layer is to represent and manipulate the semantic information provided by the Extraction layer in order to fulfil the user-oriented needs of the access layer. The primary representation requirement at this level is representing the structured information as extracted from the repositories by the extraction layer. Primary computation requirements at this level are a query-service for the structured information extracted from the documents.
WP-3.1 Representing the structured information
We expect to be able to conform to existing, recently defined or currently arising WWW-standards. Both XML and RDF have now been adopted as standards for representing the structure of documents (XML), and for representing meta-information about documents (RDF). By conforming to such existing or arising standards, the On-To-Knowledge toolkit will smoothly interface with technology developed both in other EU-sponsored projects and world-wide.
WP-3.2 Query language
For many of the services at the access layer it will be necessary to pose and answer queries on the structured information that has been obtained from the unstructured information repository. Unlike existing query languages, the vocabulary of such queries is not determined by the accidental format of the information source, but instead by the ontological categories that have been defined for the information. Examples of existing query languages are again the language from the Ontobroker and WebMaster projects (AIFB, VU, AIdministrator), but also a number of XML query languages proposed to the W3C. Queries in the On-To-Knowledge query language will typically compute values for ontological attributes of the information sources. Such ontological attribute-value pairs are subsequently used as in tasks at the access layer such as query-answering, agent-based information access and user-profile construction.
WP-3.3 Query engine
Since the On-To-Knowledge toolkit aims at enterprise wide scalability, efficient query evaluation techniques will be required for the query engine. Work has already been done at AIdministrator on query optimisation (e.g. by query transformation, and by heuristic query ordering), and on incremental evaluation strategies (minimising the re-evaluation of queries on updated information sources). Work has already been done at AIFB on exploiting data-base technology for efficient query processing.
WP-4: Developing various means for Information Access
Go to the list
The aim of this work package is to develop advanced push and pull techniques to lower the thresholds for accessing this information. Agent-based techniques as well as state-of the art querying and visualisation techniques can fully employ the formal annotations to guide user access of information. Ontologies will provide the crucial means needed to integrate the resources dispersed across web pages into a coherent corpus of interrelated information.
WP-4.1 Search Facility
The search facility will use a collection of information agents to find and filter relevant information that matches the user's interests. The information agents will search the World Wide Web, (WWW), a company's Intranet, News Groups, or organisational memories based on the user's profile.
The functionality of these information agents will go beyond existing because these agents will be capable to exploit the interrelationships between selected pieces of information. Using ontologies the agents can put otherwise isolated information into a meaningful context. The implicit structures revealed help users use and manage information more efficiently
WP-4.2 Knowledge Sharing Facility
Retrieved information can only reach its full potential value as useful knowledge if it is available to the right people at the right time. The knowledge sharing facility will ensure that this is the case, by automatically disseminating all shared information directly to those community members who need to have it. A quality information resource can build up, and can be searched at any time using the ontology-supported search facility.
View definition on ontologies: On-To-Knowledge provides support in defining personalised views on a joint ontology. Then, agents make use of a common ontology but are still able to use their internal representation by explicitly linking their internal view with the external ontology.
WP-4.3 Visualisation Facility
The visualisation facility will be a collaborative 3D-information visualisation tool for finding, retrieving and sharing information. Since humans are instinctively attuned to a 3D world, 3D interactive multimedia information sources offer the possibility of an intuitive and natural way to explore complex ideas, concepts and information. The visualisation facility aims to provide a virtual environment wherein users can meet and share relevant information.
WP-4.4 Semi-automatic user Profile construction
We will develop monitoring agents that track people's informational behaviour when accessing knowledge repositories and the underlying ontologies. This will enable such monitoring agents to build up ever more accurate profiles of their respective users. User can also define their own personalised view, their user profile and their information agents in terms of an ontology. Using this profile, agents can find, share and manage information.
WP-5: Methodology and Management Guidelines
Go to the list
The aim of the workpackage on methodology is to develop management guidelines and methods ensuring a successful introduction of knowledge management tools into an enterprise. Here, we need an integrated approach balancing the organisation and management aspects on the one hand, and the information technology and systems aspects on the other hand.
WP-5.1 Baseline version of the On-To-Knowledge KM
Very early on in the project, we will produce a baseline version of the On-To-Knowledge KM guidelines for management. They are based on CommonKADS, literature study and a systematisation of existing practical experiences.
WP-5.2 Employment and Evaluation of the baseline version of the On-To-Knowledge KM
The baseline version is employed and evaluated in the set of case studies envisaged. Here, case study partners are required to use and give feedback on the KM guidelines, and methodology partners are required to give active assistance and support the organisation of feedback.
WP-5.3 Gradual Expansion and Revision of the baseline version
The baseline version is gradually expanded and revised by the methodology partners as experiences become increasingly available covering the different issues mentioned.
WP-5.4 Final version of the On-To-Knowledge KM Guidelines
At the end of the project, a validated final version of the On-To-Knowledge KM guidelines is produced that is general and comprehensive and practically illustrated by real-life industry cases. The result will be the On-To-Knowledge methodology being described in a comprehensive guide book.
WP-6: Case study on organisational memory
Go to the list
WP-6.1 Application Study
A first, 6 month application study will yield a detailed requirements analysis of the whole case study. Where possible, already existing versions of tools will be employed to achieve as much of the intended, final functionality as possible.
WP-6.2 First Prototype
The starting point for the first real prototype are two already existing information systems for supporting office work. The first prototype aims at achieving the following two goals:
  • Obtain a smooth integration of weakly structured knowledge with formally represented knowledge.
  • Based on that integration, provide a maintenance tool that makes extension and modification of both, weakly structured and formalized knowledge easy. The tool ensures consistency by keeping track of link structure and taking into account the type and contents of the informational units being related.
WP-6.3 OM Application and Evaluation
This work package will investigate how to apply the tools and techniques developed during project to meet the requirements identified in the requirement analysis. The objective of the evaluation aspect of this task is also to measure and report the progress of the On-To-Knowledge project in terms of the five key criteria formulated by the commission. In this way, we end up with (the prototype of) a full-fledged Organizational Memory System.
While the first prototype has focused on maintenance aspects, the second prototype concentrates more on the aspects of providing an end user optimal access to the company's organizational memory.
WP-7 A case study on call centers
Go to the list
Call Centres are becoming an increasingly important mechanism for customer contact in many European industries. Work must be organised and distributed efficiently without a generic view of either customers or customer service people. Every transaction should emphasise the uniqueness of both the customer and the customer service person. To do this one needs effective knowledge management. This includes knowledge about the customer but also knowledge about the customer service person so that the customer is directed to the right person to answer their query. This knowledge must also be used in a meaningful and timely way.
WP-7.1 Requirement Analysis
This work package will look at one or more of BT's own Call Centres to identify requirements and opportunities for effective knowledge management as outlined above. More specifically, call centre agents tend to use a variety of electronic sources for information when interacting with customers including their own specialised systems, customer databases, the organisation's intranet and, perhaps most importantly, case bases of best practice.
WP-7.2 Tool Application and Evaluation
This work package will investigate how to apply the tools and techniques developed during project to meet the requirements identified in the requirement analysis. The tools are expected to provide a natural and intuitive front end onto the heterogeneous information sources used within Call Centres.
The evaluation criterion is to what degree the application of tools and methodology can ensure that the performance of the best Call Centre agents is transferred to the others. The objective of the evaluation aspect of this task is also to measure and report the progress of the On-To-Knowledge project in terms of the five key criteria formulated by the commission.
WP-8: A case study on knowledge management in a virtual enterprise
Go to the list
The aim of this case-study by Enersearch is to evaluate the On-To-Knowledge toolkit for increased knowledge transfer between different researchers in the partner's organisation, and between researchers and specialist from Enersearch's shareholding companies who have a general interest in getting up-to-date information about achieved R&D results.
Two different interest groups from the virtual organization will be selected in the case-study:
    1. researchers from different fields, and
    2. specialists from the shareholders' organizations.
WP-8.1 IT in Energy ontology design and implementation
The first task is to structure (parts of) the current web-site in a way suitable for both storing and retrieving information. This will be done in two steps:
    1. Make the domain ontology for this context with the help of the On-To-Knowledge tools and methodologies. That is, identify and describe the content identifiers and structures of this application domain in a format suitable for electronic storage and retrieval.

    2. Link this ontology to the materials on the web-site (both public and confidential parts). That is, classify the current content of the web-site using the abstractions from item 1 above, and support information users and providers with tools for taking advantage of the content-based classification approach.
This task will be coordinated with the work ongoing in the Methodology WP5, by using and giving feedback on the management and development guidelines proposed there.

WP-8.2 Preparation of the evaluation experiment
The evaluation experiment will be set up by arranging two different focus groups of users who will actually carry out the evaluation experiment. Again, these groups can mainly be characterized as information providers and information users. In addition, a scenario for the evaluation experiment will be prepared, including tools for organizing the desired feedback (interview schemes, questionnaires and the like).
WP-8.3 Application study and Evaluation
The actual evaluation experiment is carried out according to the results of Tasks 8.1 and 8.2. The results of the experiment will be laid down in a final deliverable of WP8, collecting the experiences with the On-To-Knowledge tools and methodology. The evaluation of the On-To-Knowledge tools and methods will be in the form of evaluation reports serving as feedback to the On-To-Knowledge tool developers.
WP-9: Information dissimination, take-up of results, and Exploitation
Go to the list
The goal of this task is to distribute the results of the project and convincing other organisations that it may be valuable for them to apply our techniques. We will also aim for getting feedback from external sources to improve our results.
WP-9.1 Web Site
We will run a web site of the project which may include downloadable beta versions of programs.
WP-9.2 Workshops
We will organise two workshops with external persons to discuss our results.
WP-9.3 Exploitation Activities
The project objectives represent major business opportunities for all the partners involved and coordinated exploitation activities will take place throughout the project.
WP-10: Management
Go to the list
WP-10.1: Project Management
Main responsible partner is VU. The structure and activities of project management are described elsewhere.
WP-10.2 Exploitation Management
BT will assume the main responsibility for the exploitation management. It will assist the other partners in identifying relevant market segments and potential customers.
To this end, it will provide:
  • market reports (including estimates of potential market size)
  • competitor reports,
  • technology assessments.
WP-11: Management
Go to the list
WP-11.1: OMM Analysis & Design
Main responsible partner is SAI. The structure and activities of project management are described elsewhere.
WP-11.2 OMM Development
WP-11.3 Reasoning Enhancements
WP-11.4 OMM Integration