The Internet Reasoning Service - IRS - is KMi's Semantic Web Services framework, which allows applications to semantically describe and execute Web services. The IRS supports the provision of semantic reasoning services within the context of the Semantic Web.
 
There are currently two implementations: IRS-II and IRS-III
IRS-II has been applied within the MIAKT Project .
IRS-III is currently being applied within the DIP , Super, Luisa and LHDL projects, and the WSMO Working group .

IRS Framework

The overall framework of IRS is shown in figure 1. The main components are the IRS Server, the IRS Publisher and the IRS Client. These components communicate through SOAP.



Figure 1. The IRS framework

IRS Server
The IRS server holds descriptions of Semantic Web Services at two different levels. A knowledge level description of components is represented internally in OCML. Additionally, two sets of mappings are used to connect the knowledge level descriptions to a specific Web service.

IRS Publisher

The IRS Publisher plays two roles in the IRS framework. Firstly, it links Web services to semantic descriptions within the IRS server. Secondly, it automatically generates a wrapper which allows standalone Lisp or Java code to be invoked as well as a Web service through its WSDL description.

IRS Client
A key feature of IRS is that Web service invocation is capability driven. An IRS user simply asks for a task to be achieved and the server selects and invokes an appropriate Web service.


IRS-II Approach

IRS-II follows the UPML framework developed within the IBROW project. The UPML framework partitions knowledge into ontologies, domain models, task models, and problem solving methods (PSMs) which are connected via bridges. Figure 1 shows how a single knowledge based application would be described in UPML terms. The components of figure 2 are:
  • Domain models - Domain models describe the domain of an application (e.g. vehicles, a medical disease),

  • Task models - a generic description of the task to be solved, specifying the input and output types, the goal to be achieved and pre and post-conditions,

  • Problem Solving Method - a description of the generic reasoning process to be applied, for example, heuristic classification or propose and revise.

  • Bridges - contain mappings between the different model components within an application. For example, the refinement process in heuristic classification may be mapped onto a taxonomic hierarchy of attributes within some domain.
Each knowledge model type (task, PSM and domain model) is supported by appropriate ontologies.




Figure 2. An example of how a knowledge based application would be created out of reusable components within the UPML Framework.



IRS-II movie: Patient Shipping application

QuickTime (26Mb)

Flash (2 Mb)

IRS-III movie: A Case-Study on E-Government (Change of Circumstance Scenario)

QuickTime (17Mb)

A tutorial about IRS-II and several about its successor, IRS-III, can be downloaded from KMi's DIP project webpage.

^ Back to top

IRS-III Approach

IRS-III is a platform and infrastructure for creating WSMO-based Semantic Web Services, building upon the previous implementation, IRS-II. The specific extensions in IRS-III are:

  • UPML based types of knowledge models: domain, task, problem solving method and application have been extended to include goal models, mediator models and web service models;
  • The WSMO ontology has been implemented and slightly extended in OCML;
  • A WSMO specific Java API has been created;
  • The IRS browser has been customised to reflect WSMO.


Various IRS-III tutorials can be downloaded from KMi's DIP project webpage.

The tool WebOnto can be used to visualize and edit IRS-III ontologies (in OCML)

^ Back to top


Download

The IRS has now been open-sourced under a BSD-like license. You can get it from http://github.com/kmi/irs.

Several older packages of IRS materials include:


^ Back to top

Related Publications

SQL Error:
SELECT * FROM people ORDER BY id