Download - JPL_Presentation
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 1/33
Semantic Web
&Cased Based Reasoning
AIST Meeting JPL, CA 2003
Mehmet S. Aktas
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 2/33
Outline
Semantic Web Overview
Semantic Web
Motivations Ontology Languages
Semantic Web and Cased Based Reasoning
Cased Based Reasoning Overview
Cased Based Reasoning CBR Process
Conversational Cased Based Reasoning
AIST Meeting JPL, CA 2003
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 3/33
AIST Meeting JPL, CA 2003
Semantic Web Overview
³The Semantic Web is a major research initiative of the World Wide
Web Consortium (W3C) to create a metadata-rich Web of resources
that can describe themselves not only by how they should be
displayed (HTML) or syntactically (XML), but also by the meaning of the
metadata.´From W3C Semantic Web Activity Page
³The Semantic Web is an extension of the current web in which
information is given well-defined meaning, better enabling computers
and people to work in cooperation.´
Tim Berners-Lee, James Hendler, Ora Lassila,
The Semantic Web, Scientific American, May 2001
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 4/33
AIST Meeting JPL, CA 2003
Motivations
Difficulties to find, present, access, or maintain
available electronic information on the web
Need for a data representation to enable software
products (agents) to provide intelligent access to
heterogeneous and distributed information.
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 5/33
AIST Meeting JPL, CA 2003
The Semantic Stack and Ontology Languages
XML, XML S chem a
DF
DAML,
OIL,
DAML+OIL O L L ite
DF Schem a
O L DL
O L Fu ll
From ³The Semantic Web´ technical report by PierceThe Semantic Language Layer for the Web
A
B
A = Ontology languages based on XML syntax
B = Ontology languages built on top of DF and DF Schema
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 6/33
AIST Meeting JPL, CA 2003
Resource escription ramework (R ) -
Resource Description Framework (RDF) is a framework for
describing and interchanging metadata (data describing the web
resources).
RDF provides machine understandable semantics for metadata.
This leads,
better precision in resource discovery than full text search,
assisting applications as schemas evolve,
interoperability of metadata.
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 7/33
AIST Meeting JPL, CA 2003
Resource escription ramework (R )- RDF has following important concepts
Resource : The resources being described by RDF areanything that can be named via a URI.
Property : A property is also a resource that has a name, for instance Author or Title.
Statement : A statement consists of the combination of a
Resource, a Property, and an associated value.
Example: Alice is the creator of the resource http://www.cs.indiana.edu/~Alice.
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 8/33
AIST Meeting JPL, CA 2003
The ublin Core efinition Standard
RDF is dependent on metadata conventions for definitions.
The Dublin Core is an example definition standard whichdefines a simple metadata elements for describing Webauthoring.
It is named after 1995 Dublin (Ohio) Metadata Workshop.
Following list is the partial tag element list for Dublin Corestandard.
Creator: the primary author of the content
Date: date of creation or other important life cycle events
Title: the name of the resource
Subject: the resource topic
Description: an account of the content
Type: the genre of the content
Language: the human language of the content.
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 9/33
AIST Meeting JPL, CA 2003
Example
http://www.cs.indiana.edu/~Alice
creator=
http://purl.org/dc/elements/1.1/creator
Alice is the creator of the resource http://www.cs.indiana.edu/~Alice.
� Property ³creator´ refers to a specific definition. (in this example by ublin Core
efinition Standard). So there is a structured UR for this property. This UR makes this
property unique and globally known.
� By providing structured UR we also specified the property value lice as following.
³http://www.cs.indiana.edu/People/auto/b/ lice´
Alic e
ResourceProperty
Property
Value
Inspired from ³The Semantic Web´ technical report by Pierce
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 10/33
AIST Meeting JPL, CA 2003
Example
Alice is the creator of the resource http://www.cs.indiana.edu/~Alice .
Inspired from ³The Semantic Web´ technical report by Pierce
<rdf:RDF xmlns:rdf=´http://www.w3c.org/1999/02/22-rdf-syntax-ns##́
xmlns:dc=´http://purl.org/dc/elements/1.1´
xmlns:cgl=´http://cgl.indiana.edu/people´>
<rdf:Description about=´ http://www.cs.indiana.edu/~Alice´>
<dc:creator>
<cgl:staff> Alice </cgl:staff>
</dc:creator>
</rdf:RDF>
� Information in the graph can be modeled in diff. ML organizations. Human readers would
infer the same structure however general purpose applications would not.
�Given R model enables any general purpose application to infer the same structure.
Why bother to use
RDF instead of XML?
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 11/33
AIST Meeting JPL, CA 2003
R Schema (RDFS )
RDF Schema is an extension of Resource Description Framework.
RDF Schema provides a higher level of abstraction than RDF.
specific classes of resources , specific properties,
and the relationships between these properties and other resources can bedescribed.
RDFS allows specific resources to be described as instances of moregeneral classes.
RDFS provides mechanisms where custom RDF vocabulary can bedeveloped.
Also, RDFS provides important semantic capabilities that are used byenhanced semantic languages like DAML, OIL and OWL.
It resemblesobjected-oriented
programming
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 12/33
No standard for expressing primitive data types such as integer, etc.
All data types in RDF/RDFS are treated as strings.
No standard for expressing relations of properties (unique, transitive,
inverse etc.)
No standard for expressing whether enumerations are closed.
No standard to express equivalence, disjointedness etc. among
properties
Limitations of R /R S
AIST Meeting JPL, CA 2003
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 13/33
RDF\RDFS define a framework, however they have limitations. There is a
need for new semantic web languages with following requirements
T hey should be compatible with (XML, RDF/RDFS)
T hey should have enough expressive power to fill in the gaps in
RDFS
T hey should provide automated reasoning support
Ontology Inference Layer (OIL) and DARPA Agent Markup Language
(DAML) are two important efforts developed to fulfill these requirements.
Their combined efforts formed DAML+OIL declarative semantic language.
AIST Meeting JPL, CA 2003
DAML OIL and DAML OIL - I
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 14/33
DAML+OIL is built on top of RDFS.
It uses RDFS syntax.
It has richer ways to express primitive data types.
DAML+OIL allows other relationships (inverse and transitivity) to bedirectly expressed.
DAML+OIL provides well defined semantics, This provides followings:
Meaning of DAML+OIL statements can be formally specified. Machine understanding and automated reasoning can be supported.
More expressive power can be provided.
AIST Meeting JPL, CA 2003
DAML OIL and DAML OIL - II
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 15/33
Example: T. Rex is not herbivore and not a currently living species.
This statement can be expressed in DAML+OIL, but not in RDF/RDFS
since RDF/RDFS cannot express disjointedness.
DAML+OIL provides automated reasoning by providing such expressive
power.
For instance, a software agent can find out the ³list of all the carnivores that
won¶t be any threat today´ by processing the DAML+OIL data representation
of the example above. RDF/RDFS does not express ³is not´ relationships and exclusions.
AIST Meeting JPL, CA 2003
Example
How is DAML+OIL isiff r t th RDF/RDFS?
From ³The Semantic Web´ technical report by Pierce
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 16/33
Web Ontology Language (OWL) is another effort developed by the OWLworking group of the W3Consorsium.
OWL is an extension of DAML+OIL.
OWL is divided following sub languages. OW L Lite
OW L (Description Logics) DL
OW L Full ± limited cardinality
OWL Lite provides many of the facilities of DAML+OIL provides. In
addition to RDF/RDFS tags, it also allows us to express equivalence,identity, difference, inverse, and transivity.
OWL Lite is a subset of OWL DL, which in turn is a subset of OWL Full.
AIST Meeting JPL, CA 2003
Web Ontology Language (OWL)
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 17/33
Developing new tools, applications and architectures on top of the
Semantic Web is the real challenge.
AI techniques should be used to utilize the Semantic Web up to its
potentials.
CBR is an AI technique based on reasoning on stored cases.
CBR technique can be applied to do intelligent retrieval on metadata
of codes related Earthquake Science.
From Semantic Web to Cased Based
Reasoning
AIST Meeting JPL, CA 2003
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 18/33
AIST Meeting JPL, CA 2003
CBR is reasoning by remembering: It is a starting point for newreasoning
P roblem-solving: CBR solves new problems by retrieving andadapting records from similar prior problems.
I nterpretive/classification:CBR understands new situations by
comparing and contrasting them to similar situations in the past
Case-based reasoning is a methodology of reasoning from specificexperiences, which may be applied using various technologies(Watson 98)
What is CBR?
Overview of C
ase-Based Reasoning
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 19/33
Everyday Examples of CBR
Remembering today¶s route from the place you live to campus andtaking the same route.
Diagnosing a computer problem based on a similar prior problem.
Predicting an opponent¶s actions based on how they acted under similar past circumstances
Assessing a hiring candidate by comparing and contrasting toexisting employees
Wh
at isC
BR?
AIST Meeting JPL, CA 2003
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 20/33
CBR
Process
What is a Case?
Input cases are descriptions of a specific problem.
Stored cases encapsulate previous specific
problem situations with solutions.
Another way to look at it:
Stored cases contain a lesson and a specific
context where the lesson applied.
The context is used to determine when thelesson may apply again.
AIST Meeting JPL, CA 2003
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 21/33
CBR
Process
When and how are cases used?
Given a Problem Description (P.D.) to be solved,
CBR follows a cyclical process.
REtrieve the most similar case(s)
REuse the case(s) to attempt to solve the problem
REvise the proposed solution if necessary
REtain the new solution as a part of new case.
AIST Meeting JPL, CA 2003
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 22/33
CBR
P
rocessProblem
etrieve
euse
evise
etain
Proposed solutionConfirmed solution
Case-Base
T e CB Cycle
AIST Meeting JPL, CA 2003
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 23/33
Conversational CBR (CCBR) CCBR is a method of CBR where user interacts
with the system to retrieve the right cases.
System responds with ranked cases and
questions at each step
Question-answer-ranking cycle continues until
success or failure
AIST Meeting JPL, CA 2003
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 24/33
Conversational CBR CCBR facilities
Question management facility
Case management facility
GUI for user-system interaction
Facilities to display questions or cases
AIST Meeting JPL, CA 2003
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 25/33
A Prototype CCBR Application
AIST Meeting JPL, CA 2003
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 26/33
A Prototype CCBR Application
Purpose Intelligent retrieval on metadata describing codes written for
earthquake science.
Guidance on how to run the codes to get reasonable results. Guidance for inexpert users to browse and select codes
Casebase disloc - produces surface displacements based on multiple
arbitrary dipping dislocations in an elastic half-space simplex - inverts surface geodetic displacements to produce fault
parameters
VC - simulates interactions between vertical strike slip faults.
AIST Meeting JPL, CA 2003
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 27/33
A Prototype CCBR Application
Classification Initial effort ± dummy cases created to classify the different codes
A general approach is needed
AIST Meeting JPL, CA 2003
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 28/33
A Prototype CCBR Application
AIST Meeting JPL, CA 2003
CCBR CASE
Problem Solution
¡ ¢
F
¡ t
¢ r
¡ ¢
F
¡ t ¢ r
¡ ¢
F
¡ t
¢ r
= <Question Answer>
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 29/33
A Prototype CCBR Application
How does Case Ranking take place in CCBR? Retrieved cases are sorted based on their consistency
with the query case.
As the questions are answered more cases areeliminated.
A case is ruled out only if there is a conflict between thecase and the query case
Consistency number for a case remains same if the case
has no answer for the question. Consistency number for a case gets incremented if the
case has the same answer to the question as the querycase.
AIST Meeting JPL, CA 2003
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 30/33
A Prototype CCBR Application
AIST Meeting JPL, CA 2003
CCBR CASEBASE
£ Cas £
Feature 1
Feature 2
Feature 5
£ Cas £ = <Problem Solution>
Feature 1
Feature 2
Feature 3
Feature 4
A Case from
CASEBASE
Query Case
IF ((A.Feature1.Solution =B.Feature1.Solution) &
(A.Feature2.Solution = B.Feature2.Solution))
THEN Consistency # = 2
A B
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 31/33
A Prototype CCBR Application
How does question ranking take place in CCBR?
Questions can be ranked based on their frequency factor
Questions can be ranked based on predefined inferencerules
Only distinguishing questions are to be ranked
Questions can be YES/NO questions, multiple choice
questions or questions with numerical answers.
AIST Meeting JPL, CA 2003
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 32/33
W3C Semantic Web Activity Page. Available fromhttp://www.w3.org/2001/sw/.
T. Berners-Lee, J. Hendler, and O. Lassila, ³The Semantic Web.´Scientific American, May 2001.
Resource Description Framework (RDF)/W3C Semantic Web ActivityWeb Site: http://www.w3.org/RDF/.
D. Brickley and R. V. Guha (eds), ³RDF Vocabulary DescriptionLanguage 1.0: RDF Schema.´ W3C WorkingDraft 23 January 2003.
The DARPA Agent Markup Language Web Site: http://www.daml.org.
OIL Project Web Site: http://www.ontoknowledge.org/oil
References
AIST Meeting JPL, CA 2003
8/7/2019 JPL_Presentation
http://slidepdf.com/reader/full/jplpresentation 33/33
References CBR on the web
http://www.cbr-web.org
Case-Based Reasoning Resources
http://www.aaai.org/Resources/CB-Reasoning/cbr-resources.html
AI Topics - CBR
http://www.aaai.org/AITopics/html/casebased.html
A mailing list including announcements, questions, and discussion about
CBR, managed by Ian Watson [email protected]
Riesbeck & Schank, Inside Case-Based Reasoning, Erlbaum, 1989.
Kolodner, Case-Based Reasoning, Morgan Kaufmann, 1993.
AIST Meeting JPL, CA 2003