A Logical Framework for Web Service Discovery

1
A Logical Framework for Web Service Discovery

• The Third International Semantic Web Conference
• Hiroshima, Japan, 08-11-2004
• Michael Kifer1, Rubén Lara2, Axel Polleres2,
  Chang Zhao1,
• Uwe Keller2, Holger Lausen2, and Dieter Fensel2
• ruben.lara_at_deri.org
• 1Department of Computer Science University at
  Stony Brook, New York, USA
• 2Digital Enterprise Research Institute,
  Innsbruck, Austria, and Galway, Ireland

2
Overview

1. Introduction
2. Proof obligations and formalization
3. Realization
4. Semantics of rule reification
5. Conclusions future work

3
Automatic discovery

• Current Web Services have to be selected and
  hard-wired at design time
• No dynamic reconfiguration of services
• Semantics can enable the automatic location of
  Web Services providing particular functionality

4
WSMO WSML
Objectives that a client may have when
consulting a Web Service

• Semantic description of Web Services
• Capability
• Interfaces

Provide the formally specified terminology used
by all other components
Connectors between components to bypass
heterogeneity
F-Logic Transaction Logic
5
The problem

• Matching capabilities of existing Web Services
  against the goal described by the requester
• Consideration of the functionality of the Web
  Service
• Distinction between discovery and contracting
• Example implementation using

6
Overview

1. Introduction
2. Proof obligations and formalization
3. Realization
4. Semantics of rule reification
5. Conclusions future work

7
Goals, capabilities mediators

• Goal describes (in terms of domain ontologies)
  the desired state of
• Information space
• State of the world
• Web Service capabilities describe (in terms of
  domain ontologies)
• What the service expects to provide its
  functionality
• What is guaranteed to hold after execution
• wgMediators link Web Services and goals,
  resolving heterogeneity
• Resolve possible terminology differences

8
Logics and scalability issues

• Logic can be used to formalize goals,
  capabilities and proof obligations
• Scalable framework must rely on a relatively
  small number of logicians

• Mediation Provider
• - Bulk of logical expertise
• Link ontologies, not customers and providers

• Service Provider
• - modest requirements
• Capabilities written to relatively simple
  ontologies
• Relatively simple types of rules

• Customer
• - no training in KR
• pre-defined discovery queries
• goal ontology

9
Proof obligations (I)

• Set of imported ontologies O
• Goal G
• Service capability C (Ceff and Cpre)
• wgMediator wg
• takes a goal G and constructs input Inwg(G)
  suitable for services mediated
• Converts the goal into a postcondition Postwg(G)
  expressed in terms of the service ontology
• Mediation can be complex
• Goals can be expressed in a very high level
  syntax
• Service capabilities can be rather simple

10
Proof obligations (II)

• Service discovery
• Given a goal G, can the service execute in a way
  such that G can be achieved?
• Service contracting
• Given an actual input to a specific service, does
  this input lead to the results expected by the
  requester?

11
Proof obligations (III)

• Proof obligations before
• Deal with a particular service
• Different services have different effects
• Use of transaction logic
• is the sequence operator
• is the hypothetical operator

12
Overview

1. Introduction
2. Proof obligations and formalization
3. Realization
4. Semantics of rule reification
5. Conclusions future work

13
Realization

• Use of
• Support for F-Logic, HiLog, Transaction Logic and
  rule reification
• Geographic ontology

14
Realization (II)

• Goal ontology
• Service1

Conditions over the input
Input is a search -gt provide itinerary
wgMediator used
15
Realization (III)

• Service 3
• Goals

Uses goal ontology
Region!
16
Realization (IV)

• Mediator

Takes the goal and constructs input to the service
Takes the result and checks it according to the
format specified in the goal
17
Realization

• Discovery

Construct input
Generate effects
Assume effects
Check goal
Remove effects
18
Overview

1. Introduction
2. Proof obligations and formalization
3. Realization
4. Semantics of rule reification
5. Conclusions future work

19
Semantics of rule reification

• Model theory for F-Logic extended with rule
  reification defined
• Reified F-Logic avoids paradoxes through two
  restrictions
• No negation is allowed in the rule head, and
• Reification of negation of any fact or any rule
  is not permitted

20
Overview

1. Introduction
2. Proof obligations and formalization
3. Realization
4. Semantics of rule reification
5. Conclusions future work

21
Conclusions

• Logical framework and realization for
• dynamic discovery of Web Services
• verification of contractual statements
• Scalable framework in terms of human resources by
  exploiting mediators
• The framework captures the relation between
  inputs and effects, thus providing more accurate
  descriptions and discovery
• Easily extendable to include invocation

22
Future work

• In progress
• Alignment with WSML
• Integration with other types of web service
  discovery
• Further investigation on border between
  ggMediation and wgMediation
• Complete knowledge goals in the absence of
  sufficient domain knowledge
• Implementation of WSMO discovery engine
• Planned
• Integration with composition

23
Conclusions

• lt/A Logical Framework for Web Service Discoverygt
• ltQAgt