Abstract

New Web Services are being made available on the internet all the time, and while some of them provide completely new functionality, most are slight variations on already existing services. I am interested in the problem of enabling systems to take advantage of new services without the need for reprogramming them. An existing system can only make use of the new service if it has some notion as to what functionality the service provides. There are three approaches to gaining such knowledge, which I term standardization, semantic markup and induction:• define standard schemas and sets of operations (service descriptions) for the information domain and require that all service providers use those schemas.• rely on service providers to annotate their services with semantic labels, corresponding to concepts from an ontology, and then employ semantic web techniques to reason about mappings between ontologies “understood” by the client, and those used by the provider.• place no requirements on the service provider, but employ schema matching (Rahm & Bernstein 2001) and service classification techniques (Heß & Kushmerick 2003) to hypothesize what functionality the service might provide. I follow the third approach, taking the idea one step further by actively querying sources to see if the output agrees with that defined by the model. Moreover, I search not only for identical services, but see whether a new service has a different scope, or indeed combines the functionality of other known services.
I restrict the problem to that of dealing with services which only produce information, without having any affect on the “state of the world”. For …

Date

July 9, 2005

Authors

Mark James Carman, CA Knoblock

Journal

PROCEEDINGS OF THE NATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE

Volume

20

Issue

4

Pages

1640

Publisher

Menlo Park, CA; Cambridge, MA; London; AAAI Press; MIT Press; 1999