Upload
lionel-may
View
212
Download
0
Embed Size (px)
Citation preview
GODO: Goal driven orchestration for Semantic
Web Services
… or how do spells work in the XXI century
Juan Miguel Gomez, Mariano Rico, Francisco Garcia and Christoph Bussler
Digital Enterprise Research Institute
WIW-04 2
Outline
• Introduction• SWS and Goal Driven Orchestration• The GODO architecture• The travel plan use case• Future research and directions
WIW-04 3
Introduction
• For centuries, mankind has looked for a way of making their wishes come true just stating them– Ancient story of the Middle East. “Then Al – Hadin,
son of Harun Al-Raschid commanded the genius to bring him one thousand million treasures… and so he did”, The 1001 nights. Robert Graves edition.
– Middle Age: Luciano from Samosata famous mirror, which could be asked for anything on earth.
– Present: Paris FNAC example– Future: There comes the robots…
WIW-04 4
Some promises of SWS• In WSMO/ WSMX a goal represents the
wish that a client may have when he consults a web service and it also contains a list of preferences
• These preferences represent constraints on non-functional properties of a web service i.e. they narrow the scope of the selection spectrum of a web service
• WSMX promises that given a certain WSMO goal described in WSML, it can achieve it
WIW-04 5
Goal driven orchestration
• We assume for now that a goal is a single-step execution
• Orchestration is the achievement of several goals by performing all their objectives
• How can we bridge the gap between the client expressing their wishes and the achievement of them by the WSMX platform?
WIW-04 6
GODO functionality
• GODO uses Natural Language Processing techniques (e.g. Multiple classification ripple down rules) to filter the different concepts and relationships of the text to create a “lightweight ontology”
• The user writes down their goals in natural language and they are extracted from the text
• Those goals are matched and mapped to the WSMO / WSMX goals
• Those goals are sent to the WSMX
WIW-04 7
The GODO Architecture
GODO Control ManagerGUI
Language Analyzer
Goal Matcher
WSMLGoal
WSMLGoalGoal
Loader
WSMOgoal
repository
Network
Goal Sender
User wish andgoal text
Figure . The GODO architecture
WIW-04 8
The travel plan use case (I)
WIW-04 9
The travel plan use case (II)
WIW-04 10
GODO demo
• Much more fun in the demo… • Do not miss it
WIW-04 11
Future research and directions
• Main problem is the pure syntactical match from goals extracted from the text and WSML goals
• However, several useful tools out there: – The Karlsruhe TextoOnto supports semi-automatic
creation of ontologies by applying text mining algorithms
– The OntoText Knowledge Information Management (KIM) platform. KIM enables Semantic annotation of text and at more length, an automatic ontology population and open-domain dynamic semantic annotation of unstructured and semi-structured content.
• By using them it could be possible a match at a semantic level (ontologies merging and alignment techniques)
WIW-04 12
Future research and directions
• Future evolution of WSMO Orchestration will impact in our perception of orchestration so far
• We received some enthusiastic feedback from people of the cluster, so let’s expect soon GODO 2.0
Q & A