Forum e|Government - Arbeitskreis e|Gov: Strategies

Agenda and Presentations

Resume of the Workshop

• Summary of the Workshop by Josef Makolm and Silke Weiß (in English)
• Article "Semantic Help!" by Josef Makolm and Silke Weiß  in the Journal of the Austrian Computer Society (in German)
• Presentation of the Workshop given by Josef Makolm and Silke Weiß within the Special Track "Wissensmanagement in der öffentlichen Verwaltung" as part of the Triple-I (partly in German)
  

Semantic Help: an Overview

Semantic Help should improve the public service for clients, citizens and businesses. The central idea of Semantic Help is to support people by discovering their specific life situation, which usually consists of a conglomerate of different services required. Semantic Help provides a context sensitive and complete guide to the range of public services offered. It improves service localization for the users by producing individual check-lists, offering an individually configured meta-process for the orchestration of different services, or fostering Web 2.0 dynamics through which users can (partially) redesign a website according to their individual needs. Intelligent event-driven software agents can support users by guiding them through difficult procedures and by autonomous completing of incomplete data or offering pre-filled forms. Different user interfaces like avatars, sign language or just WAI conform web-interfaces are feasible.

Aim of the Workshop

The aim of the workshop is to draft a roadmap to Semantic Help by bringing together the leading scientists of the fields of ontologies, active portals and life situations as well as process orchestration:

1. to summarise the status of research and practical experience,
2. to discuss actual questions, existing obstacles and possible first implementation steps of a Semantic Help System, and
3. to identify further research challenges and objectives.

It is an additional aim of the workshop to provide compact information of the above mentioned topics, such as Semantic Help and process orchestration and ontologies in the field of knowledge and process modelling for the audience. On the other hand the audience is invited to bring in individual knowledge and ideas for discussion.

It is intended by the organisers to produce a documentation of the results, which should be published in an appropriate manner, e.g. as a book. Details will be appointed together with the workshop participants.

Abstracts of the Contributions

Automatic decomposition of (business) processes
Stefanie Betz
Petri nets have been established as a suitable language for modelling business processes with intuitive graphical notation. Furthermore, Petri nets have a mathematical foundation, which enables simulation and analysis of system behavior. To enrich them with semantic context there has been developed a petri net ontology (PNO). But, to take a step forward, we are going to present an algorithm to enable automatic combination of (business) processes, for automatically enabled (business) process interaction.

Semantic Annotation and Argument Visualisation in Content-Rich Electronic Participation Systems
Yannis Charalabidis
Abstract tba

Principles of Exchanging Legal Knowledge
Tom van Engers
While automation of administrative processes proceeds, front-office workers’ tasks become more and more knowledge intensive. Supporting the knowledge worker by providing contextual information and knowledge based support in handling cases is currently feasible thanks to Semantic Web technology, AI techniques and powerful computer systems. However realisation of support systems integrating knowledge form different domains, and taking into account contextual information about the process, case features, personal capabilities etc. may result in rather complex systems, that may be hard and costly to maintain. In my presentation I will discuss how to manage this complexity in an environment that requires adaptivity due to changing regulations and clients' demands. The presentation will address some Semantic Web principles and its consequences for an adaptive systems architecture, knowledge modelling and systems design and maintenance.

Research in Process Orchestration for Semantic Help
Marijn Janssen
A huge challenge for governments all over the world is to create customer-centric service provisioning. This requires the provisioning of ‘one-stop-shop’-services, where services of different governmental agencies are combined and integrated. Integrated governmental service-delivery has to deal with the problem of fragmentation of governments within the constitutional, legal and jurisdictional limits. Service-delivery processes include activities and/or sub-processes that are performed by various public agencies. The managing of these activities and sub-processes requires process orchestration, which can be defined as a coordinating process which runs across several service-delivery processes in different department and agencies. Public sector organizations are often criticized for failing to engage in process orchestration, as this requires conditions like clear responsibilities and agreements among agencies, the reengineering of business process, making activities and sub-processes accessible using web service technology and integration to improve semantic interoperability. As such process orchestration can be viewed as a necessary condition for customer-centric service provisioning. First, for each customer need the appropriate services and activities need to be unraveled. Next the processes provided by public and private parties need to be identified and selected. Only after knowing and understanding the chained execution, a dynamic process orchestration should be created which takes care of the execution, monitoring and control of all the activities and sub-processes in order to ensure a one-stop shop and meet the desired service levels. There is hardly any research tacking the process orchestration aspect. In this presentation I will present typical problems with process orchestration for customer-centric service provisioning and discuss the need for research into this field.

An ontology-based method for Business Process anomalies detection and correction
Michele Missikoff
Business Process (BP) modelling is a critical tasks, often involving different experts having different objectives, perspectives, and skills. BP are mainly used by business people for organization purposes, while IT experts use BP in the analysis and development of software applications. Even if we only consider business people, they are often organised in groups that see different aspects of a BP, causing a fragmented vision of such a process.
Current BP modelling tools are very good in the drawing tasks, but in terms of support to identify inconsistencies, missings, redundancies, and ambiguities, provide a limited support. Similarly, little support is provided in the search functions, when you have hundreds of diagrams and wish to perform queries on a process base. Finally, when a BP actually takes place in an organization, it is difficult to monitor it to verify if all the undertaken actions are compliant with the BP model defined.
In this work we propose a BP ontological framework aimed at supporting the development of correct BPs and the execution of correct instances. The framework, called BPAL (BP Abstract Language) is a foundational ontology, with initial axiomatization and an inference service. BPAL has been designed as a support to BPMN, the BP Modeling Notation proposed by OMG, process modeling. BPAL services are able to detect invalid BP models, to identify what axioms have been violated, and to derive the information needed to remove the inconsistencies.

Upstream knowledge design configurations to construct a user-relevant semantic help tool
Pierre Rossel & Olivier Glassey
Navigation problems increased as the number of web sites, pages and their more or less consistent cross-referencing rapidly generated a complex situation for the user of e-administrative services. Ontology-based research triggered a different pathway to confront this complexity with the claim of configuring more sensitive forms of helping interactions with the user. After a few years of such developments, new questions arise, in particular linked to the very notion of semantic performance of this type of work and the complexity of the information-knowledge relationship.
In this perspective, we have identified five main interesting approaches to the overall semantic help issue. They can be formulated as following:

1. (High-level) formalizing knowledge into explicit solutions (disambiguation and meta-levels of quickly deciding what to do best in case of doubt, always proposing choices to the user, in any case); see for that Ermine for instance;
2. Building on Wiederhold remarks (2000) that there is no global solution to the semantic barrier, browsing the field of options with multiple choice rationales, hopefully close to the user diverse view points;
3. Narrowing down the expectation, allocating ontology effectiveness only to portions of websites and functions;
4. Building on learning tools that compare lexicographic and navigation arborescence with actual usages (words, choices, sequences), trying to decrease time lapse between web design, evaluation and re-design;
5. Fostering Web2.0 dynamics or “folksonomic” patterns, in which the user, as a virtual community, can partially or totally reshape a website.

Solution 1 is very costly and heavy (defining and meta-defining a lot of things, without guarantee); solution 2 may be interesting but needs to be experimented in a variety of ways so as to detect most promising trade-offs; solution 3 may be effective, but of limited scope; it was typically a good solution for the life-event approach, but may be difficult to turn into a more universal proposal; solution 4 is effective but a bit slow, as a process, to really candidate for the status of semantic help.
Our claim is to reflect upon a methodology which would in fact build upon Web2.0 type of dynamics, but improve its relevance and effectiveness by steps, drawing as much as possible upon the other solutions’ best features. Our presentation aims at outlining a roadmap for such orientation into the semantic help problem.

Definitions

Ontology:
An Ontology consists of concepts, relations between concepts, instances and axioms. Gruber defines an ontology to be "the explicit specification of shared conceptualization, used to help programs and humans share knowledge" [1]. The conceptualisation is the formulation of knowledge about the world in terms of entities. Specification means the representation of the conceptualisation in a concrete form, by encoding it in a knowledge representation language as RDF, DAML-OIL or OWL. The objective is to create a vocabulary and semantic structure for exchanging information about a domain.

Live-Events:
A life event "is a straightforward metaphor for identifying the public services related to a specific situation that citizen often face" [3]. An application of live events is building e-governance Web portals. These portals help people, in a particular situation or event, to identify which public services they need [2].

Active-Portal:
An active Portal provides users with a complete guide through the set of necessary public services. First users have to identify the appropriate (start) life event. After the identification the active portal contacts the user who has to identify all his/her circumstances that can possibly influence the solution of the life event. As result the users should obtain a list of public services tailored to his/her individual situation. Furthermore, active portals should guide users through the set of following life events or public services triggered by the start life event [2].

Process-Orchestration:
Process orchestration means a coordinating process which runs several service-delivery processes in or across different agencies [4].

Literature / References

1. Thomas R. Gruber: A Translation Approach to Portable Ontology Specifications, Knowledge Acquisition, 1993, http://tomgruber.org/writing/ontolingua-kaj-1993.pdf, last call on the 11.03.2007.
2. Ljupčo Todorovski, Anamarija Leben, Mateja Kunstelj, Domen Cukjati, Mirko Vintar, Methology for building models of life events for active portals, in: Åke Grönlund, H. Jochen Scholl, Kim V. Andersen and Maria A. Wimmer (Eds.): Electronic Government, Communication Proceedings of the 5th International EGOV Conference 2006, Trauner Verlag, September 4-8, Krakow, Poland, ISBN: 3-540-37686-0, page 61-68, 2006.
3. Efthimios Tambouris, Konstantinos Tarabanis, Wojciech Izdebski, Mariusz Momotko, Live events revisited: representation using generic workflows, in: Åke Grönlund, H. Jochen Scholl, Kim V. Andersen and Maria A. Wimmer (Eds.), Electronic Government, Communication Proceedings of the 5th International EGOV Conference 2006, Trauner Verlag, September 4-8, Krakow, Poland, ISBN: 3-540-37686-0, page 69-76, 2006.
4. Jeffrey Gortmaker, Marijn Janssen, René W. Wagenaar, Towards Requirements for a Reference Model for Process Orchestration in e-Government, in: Michael Böhlen, Johann Gamper, Wolfgang Polasek and Maria A. Wimmer (Eds.) Proceedings: E-Government: Towards Electronic Democracy, International Conference, TCGOV 2005, Springer Verlag, Bolzano, Italy, ISBN: 3-540-25016-6, page 169-179, March 2005.