estrella

CHALLENGES

The European Project Estrella

The complexity and quantity of legislation and regulations is increasing at an alarming rate. While to a modest degree this trend can be countered by appropriate deregulation, the elimination of outdated laws and efforts to reduce bureaucratic red tape, the trend towards ever more complex laws reflects necessarily the increasing complexity of modern technology and business relationships in a global economy, and increasing public interest demands in areas such health, safety and security. This complexity causes severe problems for businesses and citizens. For example, bureaucracy costs currently make up between 2 and 5 percent of the German Gross National Product and have risen more than 25% in the last eight years. In the Netherlands the total administrative costs is 16 billion per year, including 4 billion administrative cost for tax-related obligations1. The bureaucracy costs for small and medium size companies currently reduce their profits by 50%. For citizens the risks are even greater. All citizens are expected to know and abide by the law. To the extent citizens are not able to understand their rights and obligations, they will lose faith and trust in our political institutions and the legitimacy and credibility of democracy will suffer. Legal knowledge systems are the key technology for managing the complexity of legislation and enable new online e-government services in which determinative transactions requiring the application of complex legislation, such as tax benefits or social security administration, are processed online, giving citizens and businesses immediate access to a personalized and transparent assessment of their rights and obligations. Three companies in the ESTRELLA consortium, knowledgeTools, RuleBurst and RuleWise, are the European leaders in the market of legal knowledge systems. Although they have a proven track record of successfully developing and deploying systems for public administration which enable citizens and businesses to cope with complex legislation, the further acceptance and dissemination of legal knowledge technology, and thus the further growth of this market, is severely hampered by the lack of an open platform. Public administration is understandably reluctant to invest heavily in closed, proprietary solutions.

ESTRELLA will solve this problem by bringing experts and representatives of all stakeholders together to collaborate on the development of an innovative open platform for legal knowledge systems that advances the state of the art, enabling new services, while protecting investments by being interoperable with existing commercial systems.

top

GOALS

The main technical objectives of the Estrella project are to develop a Legal Knowledge Interchange Format (LKIF), building upon emerging XML-based standards of the Semantic Web, including RDF and OWL, and Application Programmer Interfaces (APIs) for interacting with legal knowledge-based systems. To achieve and demonstrate vendor neutrality and independence, translators between the LKIF format and the existing proprietary formats of LKBS vendors participating in the project will be developed.

ESTRELLA will design a document management system for legal sources, based on a uniform XML format, enabling search, explicating relations with other documents, drafting of new documents and version management. The ESTRELLA document management system should enable two-way links to be maintained between the rules of legal knowledge systems and the legal sources upon which the rules of the model are based. This is important for providing access to relevant legal sources needed to answer questions during interactive dialogs with the system, for improving the transparency and comprehensibility of the explanations and arguments produced by the system and, last but not least, to make it feasible to efficiently maintain the knowledge base as the law changes.
This is not a trivial task. Legal sources come in many forms and formats, from different jurisdictions, in different languages, with different internal structures, etc. Legal sources are not self-contained entities but are meaningfully related to each other in one jurisdiction by for instance temporal relations, and relations of priority. Each jurisdiction has separate rules for assignment of identity, for versioning, for establishment of validity and applicability etc. These rules have to be captured in a single framework general enough to model the legal content management requirements of multiple jurisdictions, while manipulating all legal sources in a single XML format, like MetaLex. An application used to apply regimes from multiple jurisdictions would make use of existing mature national XML based services for valid legislation like the Norme in Rete service in Italy, or ADW in the Netherlands, and translate it to an appropriate XML format for internal manipulation.

To demonstrate and validate the Estrella platform, European tax related legislation and national tax legislation of two European countries will be modelled and used in the pilot applications. The finance ministries or tax administrations of several other European countries will take part in an Observatory Board to ensure generality of the approach.

top

TECHNOLOGY

Legal Knowledge Interchange Format (LKIF), building upon emerging XML-based standards of the Semantic Web, including RDF and OWL, and Application Program Interfaces (APIs) for interacting with legal knowledge systems. The LKIF will apply the state of the art in the field of Artificial Intelligence and Law, taking into account business and application requirements. Existing Semantic Web initiatives are aimed at modelling concepts (OWL "ontologies") and rules (RuleML and SWRL). The LKIF will build on but go beyond this generic work to allow further kinds of legal knowledge to be modelled, including: meta-level rules for reasoning about rule priorities and exceptions, legal arguments, cases and case factors, values and principles, and legal procedures. In addition, an OWL ontology of basic legal concepts, such as obligations, permissions, rights and powers, will be developed, which can be reused when modelling a specific legal domain, such as tax law.

A reference "inference engine" and run-time environment capable of processing knowledge bases using all the features of the LKIF will be implemented and validated in the pilot applications.

Open, general, jurisdiction independent and language independent XML based representation language for legal sources that can easily be used and coupled to LKIF knowledge models of these sources. Furthermore, we will specify and build a prototype of a Legal Content Management System. based on this representation format, using Open Source components where possible. A central requirement of automated systems based on representations of legal knowledge is an explicit link to the original legal sources the knowledge is based on. Preferably this link should be isomorphic; Each argument should be backed with plausible references to legal sources proving the validity of the argument. This is not a trivial task. Legal sources come in many forms and formats, from different jurisdictions using different legal systems, in different languages, with different internal structures, etc. Legal sources are not self-contained entities but are meaningfully related to each other in one jurisdiction by for instance temporal relations, and relations of priority and relative legislative competence. Each jurisdiction has separate rules for assignment of identity, for versioning, for establishment of validity and applicability etc.

Jurisdiction specific rules are captured in an XML based representation language specifically

designed for configuring a Content Management System, so that it can use the appropriate regime for managing legal sources from that jurisdiction. The legal sources themselves are stored in a single Content Management System architecture that is general enough to respect the legal content management requirements of multiple jurisdictions, while allowing manipulation of all legal sources in a single XML language designed for that purpose: MetaLex XML (http://www.metalex.nl). The Content Management System can use existing mature national XML based services for valid legislation like the Norme-in-Rete service in Italy (http://www.normeinrete.it), or AWB in the Netherlands (http://wetten.overheid.nl ), and translate other XML languages to the appropriate XML language for internal manipulation.

top

PARTNERS

The partners in project ESTRELLA (European project Standarlized Transparent Representations in Order to Extend Legal Accessibility) are:

• University of Amsterdam

• University of Bologna

• University of Liverpool

• Fraunhofer Fokus

• RuleWise B.V.

• RuleBurst (EUROPE) LIMITED

• KnowledgeTools International GmbH

• Interaction Design Ltd.

• SOGEI- Società Generale d’Informatica S.p.A.

• Centro nazionale per l’informatica nella pubblica amministrazione ( CNIPA)

• Hungarian Tax and Financial Control Administration

• Budapesti Corvinius Egyeetem- Budaperst

• Bundesministerium der Finanzen

• Ministerio dell’Economia e delle Finanze

• Consorzio Pisa Ricerchen SCARL