content

IEA/AIE’08IEA/AIE’08

Smart Knowledge Management

Platform:

Toward Development and Application

of Decisional DNAE Szczerbicki, C Sanin, C Toro, J Posada, J Vaquero


ContentContent

1. Knowledge

2. Knowledge Representation

3. Set of Experience

4. A Shareable Set of Experience: Decisional DNA

5. Ontology

6. Application


Knowledge Knowledge and and

Set of ExperienceSet of Experience


KnowledgeKnowledge Decision-making is supported by information. Knowledge is considered a valuable possession of

incalculable worth. Knowledge seems to be the only true source of a nation’s

economical and political strength, as well as, the key source of competitive advantage of a company (Drucker, 1995).

The means and the ability of acquisition of knowledge, through efficient transformation of information, can make the difference between the success and failure of an organization in the competitive environment of global economy and knowledge society.


Managers Managers entered into Knowledge entered into Knowledge

Administration.Administration.

They They want technologies that facilitate want technologies that facilitate

control of all forms of knowledge.control of all forms of knowledge.


KnowledgeKnowledge

Knowledge is “the fact or condition of knowing something with familiarity gained through experienceexperience or associationor association” (Merriam-Webster Dict. 2004).

“Knowledge refers to information that enables action and action and decisionsdecisions, or information with direction” (Becerra-Fernandez and Gonzalez, 2004).

Knowledge “originatesoriginates and is appliedis applied in the mind of knowers” (Mitchell 2003).

Lin et al. (2002) describe the concept of knowledge as an organized mixture of dataorganized mixture of data, integrated with rules, operations, and procedures, and it can be only developed through experienceexperience and practicepractice.


““The only source of The only source of

knowledge is experience.”knowledge is experience.”

Albert Einstein (1879 - 1955)Albert Einstein (1879 - 1955)


Knowledge Representation As knowledge should be collected, we have to represent it

in some form. One of the most complicated issues about knowledge is its

representation. Representing knowledge determines how knowledge is acquired and transformed from tacit knowledge to explicit knowledgeexplicit knowledge.

A Knowledge Representation (KR) is a substitute substitute for a thing itself - formal decision events.

A KR contains simplifying assumptionssimplifying assumptions and, possibly, extra elementsextra elements.


Knowledge Representation (KR)

The most generalized techniques of KR use logic, rules, or frames.LOGICLOGIC implicates understanding the world in

terms of individual entities and associations between them.

RULE-BASEDRULE-BASED systems view the world in terms of attribute-object-value and the rules that connect them.

FRAMESFRAMES comprise thinking about the world in terms of prototypical concepts.


What to represent?…

** OBSERVED STATISTICS REPORT for scenario TVANIM **

Label Mean Standard Number of Minimum Maximum

Value Deviation Observations Value Value

TIME IN SYSTEM 26.956 34.643 83 8.202 170.770

** FILE STATISTICS REPORT for scenario TVANIM **

File Label or Average Standard Maximum Current Average

Number Input Location Length Deviation Length Length Wait Time

1 QUEUE INSP 0.863 0.873 4 1 4.060

2 QUEUE ADJT 1.610 1.307 4 1 51.526

** ACTIVITY STATISTICS REPORT for scenario TVANIM **

Average Prob Pr. Idle Av. Service Av. Time Av Wait of Av Time Time between Probability Cumulative ProbabilityWait Time Waits Server time Bt. Arrivals those who waitin System Arrival 0

0 0 0 2 0 - 2 1 0.125 0.1250 0 0.5 2 6 0 2 2 0.125 0.250 0 0.5 1.6666667 4.5 0 1.6666667 3 0.125 0.3750 0 0.6470588 1.5 5.3333333 0 1.5 4 0.125 0.50 0 0.6538462 1.8 5.75 0 1.8 5 0.125 0.6250 0 0.6875 1.6666667 6.2 0 1.6666667 6 0.125 0.750 0 0.6216216 2 5.5 0 2 7 0.125 0.875

0.25 0.125 0.575 2.125 5 2 2.375 8 0.125 10.3333333 0.2222222 0.5348837 2.2222222 4.875 1.5 2.5555556

0.4 0.3 0.4893617 2.4 4.6666667 1.3333333 2.8 Service time Probability Cumulative Probability0.6363636 0.3636364 0.4509804 2.5454545 4.4 1.75 3.1818182 0

1 0.4166667 0.4259259 2.5833333 4.1818182 2.4 3.5833333 1 0.1 0.11.2307692 0.4615385 0.3965517 2.6923077 4.1666667 2.6666667 3.9230769 2 0.2 0.31.2142857 0.5 0.3709677 2.7857143 4.3846154 2.4285714 4 3 0.3 0.61.2666667 0.5333333 0.3650794 2.6666667 4.2857143 2.375 3.9333333 4 0.25 0.85

1.1875 0.5 0.3913043 2.625 4.4666667 2.375 3.8125 5 0.1 0.951.1176471 0.4705882 0.369863 2.7058824 4.3125 2.375 3.8235294 6 0.05 11.0555556 0.4444444 0.3846154 2.6666667 4.4705882 2.375 3.7222222

1 0.4210526 0.3902439 2.6315789 4.4444444 2.375 3.63157891 0.45 0.3764706 2.65 4.2631579 2.2222222 3.65

Multiple applications

result in decisions in

different environments,

producing formal

decision events, each

with different elements.

We have defined four

basic components that

surround decision-

making events – formal

decision events.


Formal Decision EventsThe four components are variables, functions, constraints, and variables, functions, constraints, and rulesrules, and constitute the basis for the knowledge structure.

U V

G = blue

B = high

K = average

H = good

V = 8451.54

W = 1.5

Y = 210

Z = 0.78

X = 100

VtVl

If X>70 then K = good

If W<2 then Z = 2

If G=blue then B = high

R

Max P=3X-2Y+RQ

Max K=Excellent

Min C=YQ AND B=high

F

RtRlRtl

FtFlFtl

2X+3Y-V <= 3450

H>=Excellent

G<>blue AND Y+70X<2500

C

CtClCtl


Graphic idea:

X Y W Z V K H G B …

F1 F2 F3

C1 C2 C3

…

…

R1 R2 R3 …

Set of Experience - SOESet of Experience - SOE

Set of Experience Ei = (Vi, Fi, Ci, Ri)


Variables

Rules

Constraints

Functions

SOE comprises a series of mathematical concepts (a a logical componentlogical component), together with a set of rules (a ruled based a ruled based componentcomponent), and built upon a specific event of decision-making (a frame componenta frame component).

Set of Experience - SOESet of Experience - SOE


Set of Experience - SOE It is a structure that can be used for multiple

technologies performing formal decision events. Uniqueness Adaptable Dynamic It assists at:

reducing information restrictions,developing knowledge and applying it, and implementing technologies that act as knower

and decider.


X1 X2 PromotionalChance

Status ofPayment

WorkingCondition

Min Payment Level=3X1 + 2X2

Competitor’sPayment level

PaymentLevel

6X1+X2>=21 X1+2X2>=20 X1>=0 X2>=0

IF Payment Level>=Competitor’s

Payment Level THEN Status of Payment=COMPETITIVE

Worker’s Morale >= GOOD

Worker’sMorale

IF Firing<= 1.2*Competitor’s Firing

THEN Status of Firing=VERY GOOD

Firing Competitor’sFiring

Status ofFiring

IF Promotional Chance=EQUAL THEN

Status of Promotion=VERY GOOD

1. Competitor’s payment Level = $302. Working Condition = GOOD3. Firing = 104. Competitor’s Firing = 145. Promotional Chance = EQUAL6. X1 = 27. X2 = 9

IF Working Condition>=GOOD &Status of

Payment=COMPETITIVE & Status of Firing=VERY GOOD & Status of

Promotion=VERY GOOD THEN Worker’s Morale=VERY GOOD

Status ofPromotion

RESULTS1. X1 = 52. X2 = 7.53. Payment Level = $304. Status of Payment = COMPETITIVE5. Status of Firing = VERY GOOD6. Status of Promotion = VERY GOOD7. Worker’s Morale = VERY GOOD

Constraints

Functions

Variables

Rules

Example of SOE


Applicable and Applicable and Usable SOEUsable SOE


The Set of Experience works with formal decision

events from multiple applications. All of them

having different languages, formats and

structures, and therefore, being an obstacle to the

continuous flow of information and knowledge.

A unique language facilitates the integration

practice.

A Shareable SOEA Shareable SOE


Web technologies have developed several tools for

integration of disparate systems and distributed

applications, including standards or protocols.

Languages with a defined vocabulary, structure, and

constraints for expressing information and knowledge.

Standards aim for a common language.

A Shareable SOEA Shareable SOE


4ML ABML ACAP ACML ACS X12 ADML AECM AFML AGML AHML AIF AIML AL3 AML ANATML ANML ANNOTEABGML ANZLIC APML APPEL APPML AQL ARML ASML ASTM ATML AWML AXML BannerML BCXML BEEP BGML BHTML BIBLIOML BIOML BIPS BizCodes BLM XML BML BPML BRML BSML CaseXML CaXML

CBML CDA CDF CDISC CELLML CFML ChessGML ChordML ChordQL CIDS CIDX CIM CIML CLT CML CNRP Coins ComicsML Covad xLink CP eXchange CPL CSS CVML CWMI CXML CycML DaliML DAML DaqXML DAS DASL DCMI DDI DeltaV DESSERT DIG35 DLML DML DMML DMTF DocBook DocScope DoD XML DOI

DPRL DRI DSD DSML DTB DXS EAD eBIS-XML ebXML ECML eCo EcoKnow eCX ECIX edaXML EML EMSA eosML ESML ETD-ML FieldML FINML FITS FIXML FLBC FLOWML FPML FSML GAME GBXML GDML GEDML GEML GEN GeoLang GIML GML GXD GXL GXML HEML HITIS HRMML HR-XML

HTML HTTP-DRP HTTPL HumanML Hy XM HyTime ICE ICML IDE IDML IDWG IEEE DTD IFX IML IMPP IMS Global InTML IOTP IRML IXML IXRetail JabberXML JDF JDox JECMM JigXML JLife JScoreML JSML KBML LACITO LandXML LEDES LegalXML Life Data LitML LMML LogML LTSC XML MAML MathML MatML MBAM MCF

MDDL MDSI-XML Metarule MFDX MISML MIX ML MML MMLL MoDL MOS MPML MPXML MRML MSAML MTML MusicXML NAA Ads NAML Navy DTD NewsML NFF NISO DTB NITF NLMXML NML NuDOC NVML OAGIS OAMS OBI OCF OCS ODF ODRL OeBPS Office XML OFX OIL OIM OLifE OML ONIX DTD OODT

OOPML OpenMath OPML OPX OSD OTA P3P PARLML PCIS PDML PDX PEF XML PetroML PGML PhysicsML PICS PML PMML PNG PNML PrintML PrintTalk ProductionML PSI PSL QAML QML QuickData RBAC RDDl RDF RDL RecipeML RELAX RELAX NG REPML ResumeXML RETML REXML RFML RightsLang RIXML RoadmOPS RosettaNet PIP

RSS RuleML SABLE SAE J2008 SAML SBML Schemtron SDML SearchDM-XML SGML SHOE SIF SMBXML SMDL SMI SMIL SML SMML SOAP SODL SOX SpeechML SPML SSML STEP STEPML STML SVG SWAP SWMS SyncML TalkML TaxML TDL TDML TEI ThML TIM TML TMML TMX TP TPAML

TREX TxLife UBL UCLP UDDI UDEF UIML ULF UML UMLS UPnP URI/URL UXF vCalendar vCard VCML VHG VIML VISA XML VML VMML VocML VoiceXML VRML WAP WDDX WebDAV WebML WeldingXMLXGM WellML Wf-XML WIDL WITSML WML WorldOS WSIA WSML XACML XAML XBL XBN XBRL

XBEL xCBL XCES XCFF Xchart xCIL xCML Xdelta XDF XForms XGF XGL XHTML XIOP XLF XLIFF XLink XMI XML XML Court XML EDI XML F XML Key XML MP XML News XML P7C XML Query XML RPC XML Schema XML Sign XML TP XML XCI XMLife XMLVoc XMSG XMTP xNAL XNS XOL XSBEL XSIL XUL

AIMLAIML

HTMLHTML

LPFMLLPFML

MathMLMathML

OptMLOptML

PMMLPMML

RDFRDF

RuleMLRuleML

RFMLRFML

SGMLSGML

SNOMLSNOML XMLXML

XRMLXRML

StandardsStandards


A Shareable Set of ExperienceA Shareable Set of ExperienceAmong all of the languages, XML was chosen because:

Simplicity,

Transmits not just format, but also meaningful information,

Easy to understand, read, and write,

Allows to structure information and knowledge from a graph

structure as labelled trees,

Permits defining restrictions for the document (XSD),

Supported by the W3C as a language for knowledge

representation, and

XML is the leader method for application integration.

<?xml version="1.0" encoding="UTF-8" standalone="no" ?>



-<set_of_experience xmlns: xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:noNamespaceSchemaLocation="set of experience

model.xsd">

<date>2004-11-11</date>

<hour>14:10:00</hour>

<creation>

<application> Excel </application>

<application> System </application>

<filename> payroll.xls </filename>

<filename> payroll.ces </filename>

<comment> Example of set of experience </comment>

<comment> Company Expert System </comment> </creation> <category>



<area>Human Resources</area>

<subarea>Salary Office</subarea>

<subject>Payment Level</subject>

<subject>Worker's Morale</subject>

</category>

Set of experience knowledge structure is able to be implemented in XML.

A Shareable Set of ExperienceA Shareable Set of Experience


Extending SOE Extending SOE as a KRas a KR


SOE Additional Considerations

The knowledge structure should be re-evaluated

or adjusted by users.

It is the day-to-day experience and generation of

formal decision events that make the knowledge

structure more accurate.

The knowledge structure should acts as a trained

element of “life” - Genetic History.


Extending SOE as a KR Experience is acquired while decision-making is

executed. Thus, new knowledge is produced while solving problems.

It can be compared to the process of construction of the psychological spacepsychological space of an organization.

Based upon Kelly’s theory of psychological space, we develop a knowledge structure to administer formal decision events, a structure that builds up this space with formal decision experiences. Then, this psychological space can be used for future decision-making processes based upon previous decision events.


Variables

Rules

Constraints

Functions

Image credit U.S. Department of Energy Human Genome Program (http://www.ornl.gov/hgmis).

Gene provides a Phenotype

Categorized Chromosomes

Extending SOE as a KR

Each SOE provides a value

Categorized according to type of decision


Set of Experience Set of Experience Knowledge StructureKnowledge Structure


Let C be a subset of the universe U of sets of experience named context set. A boolean expression θ containing one or many restrictions on elements of the head of the set of experience is called a selection condition on sets of experience, e.g. θ = (area = “Human Resources” AND subarea = “Salary Office“ AND aim function = “Payment Level”).

Similarity Metric of SOEKS


Similarity Metric of SOEKSSimilarity Metric of SOEKS


Because each set of experience has different elements that comprise it and each element has its own characteristics, similarity is examined separately for each of the elements; afterwards, a unique similarity measure is offered by combining their separate results.

Similarity Metric of SOEKS

VARIABLES

FUNCTIONS

CONSTRAINTS

RULES


Qualitative variables are prearranged.

Euclidean Metric with normalization.

Similar = 0, Non-similar = 1.

Similarity Metric of SOEKS: Variables

ji

n

k jkik

jkik

kV sskss

sswS

2/1

12

22

),max(


Similarity Metric of SEKSSimilarity Metric of SEKS

S1 = 0.68

S2 = 0.31

S3 = 0.74

S4 = 0.26

S5 = 0.53

S6 = 0.12

S7 = 0.07S7 = 0.07


Groups of SOE by area are:

Decisional ChromosomesDecisional Chromosomes Groups of chromosomes are =

DECISIONAL DNADECISIONAL DNA

Variables

Rules

Constraints

Functions

Variables

Rules

Constraints

Functions

Variables

Rules

Constraints

Functions

Variables

Rules

Constraints

Functions

Variables

Rules

Constraints

Functions

Variables

Rules

Constraints

Functions

Variables

Rules

Constraints

Functions

Variables

Rules

Constraints

Functions

Variables

Rules

Constraints

Functions

Variables

Rules

Constraints

Functions

Variables

Rules

Constraints

Functions

Variables

Rules

Constraints

Functions

Variables

Rules

Constraints

Functions

Variables

Rules

Constraints

Functions

Variables

Rules

Constraints

Functions

AREA 1 (marketing)AREA 1 (marketing) AREA 2 (finances)AREA 2 (finances) AREA 3 (design)AREA 3 (design)

Extending SOE as a KR


Decisional DNADecisional DNA

FFjj

RRkk

RRkk

RRkk

RRkk

RRkk

RRkk

RRkk

RR kk RRkk RRkkRRkk

RRkk RRkk RRkkRRkk

RRkkRRkk

RRkk

FFjj

FFjj

FFjj

FF jj FF jj FF jj

FFjj FFjj

CCtt

CCtt

CCtt

CCtt CC tt

CC tt

CCtt

RRkkRRkk RRkk

FFjj FFjj

CCtt

CCtt

RRkk RRkk

FFjj FFjj

RRkk

FFjj

RRkk

FFjj FFjj

RRkk

FFjj

RRkk

CCtt

RRkk

CCtt

FFjj

RRkkRRkkRRkkVV

ii

VVii

VVii

VVii VVii VVii VVii VVii

FFjj

VVii

CCtt

VViiVViiVViiVV iiVV ii VVii

VVii

VVii

VV ii

VVii

VVii

VVii

VVii

VVii

CCtt

CCttCCtt

RRkk

VVii

CCtt

FFjj

CCtt

FFjj

RRkk

CCtt

VVii

FFjj

CCtt

RRkk

VVii

FFjj

CCtt

FFjj

CCtt

FFjj

CCttCCtt

RRkkRRkkVVii VVii

FFjjFFjj

CCttCCttCCtt

FFjjFFjj

RRkk VViiVVii

CCttCCttCCttCCttCCttCC ttCC tt

CC ttCC ttCC ttCC ttCC ttCC tt

CC tt

CC tt

CCtt

CCtt

CCtt

CCtt

CCtt

FFjj

FFjjFFjj

FFjj

FF jjFF jj

FF jjFF jjFF jj

FF jj

FF jj

FF jjFF jj

FF jjFF jj

FFjj

FFjj

FFjj

FFjj

FFjj

FFjj

FFjj

FFjj

RRkk RRkkVVii VViiVVii

RR kk

VV ii

RR kkRR kk

RRkk

VV ii

VVii

Decisional Gene or SOEKS

Decisional Chromosome



It keeps:

the decisional history of a company,

the decisional experience of a company,

information that could be read by future generations

of decision makers,

knowledge in an explicit way, and

knowledge that could be shared and commercialized.



Decisional Decisional DNADNA

Ontology-Ontology-basedbased


OntologyOntology

In philosophy:In philosophy: It is the most fundamental branch of metaphysics. It studies being or existence. Tries to find out what entities and what types of entities

exist.

In Computer Science:In Computer Science: It is the explicit specification of a conceptualization. It is a description of the concepts and relationships in a

domain.

(Tom Gruber’s widespread accepted definition)


Ontology-based TechnologyOntology-based Technology

It is the field in which computer-based semantic tools and

systems are developed

Main focus is in information sharing and knowledge

management for querying and classification purposes.

Several domains of application: medical, chemical, legal,

cultural, etc.

Commonly used in AI and KR.


Computer programs can use ontologies for a variety of

purposes: inductive reasoning, classification, and

problem solving techniques are the most common

Communication and sharing of information among

different systems

Emerging semantic web systems use ontologies for a

better interaction and understanding between different

agent web-based systems

Ontology-based TechnologyOntology-based Technology


Modeling ontologiesModeling ontologies

Ontologies can be modelled using several languages; RFD and OWL are both expressed in eXtensible Markup Language-XML

OWL (Ontology Web Language) is a W3C Recommendation.

OWL facilitates machine interpretability of web content by providing additional vocabulary along with formal semantics.

Set of experience ontology-basedSet of experience ontology-based can be a scenario for exploitation of semantic data.


SOE ontology-basedSOE ontology-based

From the SOE in XML, an Ontology

modelling process was performed using

the Protégé editor (publications available)

Relationships among the different classes

of the Ontology can be seen using a plug-

in for Ontology visualization.


Modeling Set of Experience Ontology-based

Relationships among the different classes of the Ontology can be seen using a plug-

in for Ontology visualization.


Growing SystemGrowing System

Decisional DNA is shared

in this system

It is Community of

Practice distributing

knowledge

Development based on

ontology web technology

e-Decisional Communitye-Decisional Community

Co-operative Supply Chain of Knowledge

Company(Platform)Government Customers

Providers

Competence


To sum up: Platform Process

Knowledge-base layer

Integration layer

DIAGNOSISDIAGNOSIS

ExternalAnalyzer layer

Risk Analyzer layer

Si1=Li1

Si2=Li2. . .

Sim=Lim

InternalAnalyzer layer

M1

M2. . .

Mn

Se1=Le1 Se2=Le2

.

.

.

Ser=Ler

PROGNOSISPROGNOSIS

Li2 (). . .

Lim ()

Le1 ()

Le2 ()

.

.

.

Ler ()

Li1 ()

PRIORITIES

BS ()

SOLUTIONSOLUTION

Experience Creator

BS1()

BSk()

.

.

.

I1(BSi)(

Is(BSi)(

.

.

.

Intuition Creator

Ruler Creator

R1(BSi,Vj)

Rq(Bsi,Vj)

.

.

.

KNOWLEDGEKNOWLEDGE


Proposed Platform…

Understands that Knowledge is NOW an

Administrative and Technological Matter

Reduces Information Restrictions

Develops Knowledge and Applies it

A System which is Knower and Decider

Technology able to capture and store formal

decision events as explicit knowledge


ImplementationsImplementations Australia: From LIMS to LKMS Priority Priority

Research Centre for EnergyResearch Centre for Energy, Decisional DNA for Smart Use of Energy

Spain: SEMTEK, VicomTechVicomTech SOEK and Decisional DNA for industrial maintenance

Poland: GUT, Decisional DNA for banking sector

.. coming evaluation of IBM IT applications: Rational Unified Process, Method Composer, and Portfolio Manager.


Where to from Where to from here?here?


DATA

INFORMATION

KNOWLEDGE

?


Reflexive Ontologies


Reflexive Ontologies

Reflexivity addresses the property of an abstract structure of a knowledge base (in this case, an ontology and its instances) to know about itselfknow about itself.

When an abstract ontology knowledge structure is able to maintain, in a persistent manner, every query performed on it, and store those queries as individuals of a class that extends the original ontology, it is said that such ontology is reflexive.

““A Reflexive Ontology is a description of the concepts and A Reflexive Ontology is a description of the concepts and the relations of such concepts in a specific domain, the relations of such concepts in a specific domain, enhanced by an explicit self contained set of queries over enhanced by an explicit self contained set of queries over the instances”the instances”


Decisional Trust


Decisional Trust

Decisional Trust is based on three elements:

1.1. Decisional DNADecisional DNA

It offers adaptability on gathering, storing and gathering, storing and

managing decisional knowledgemanaging decisional knowledge.

It supports diversediverse decisional elements at all levels.

Decisional DNA has proven to be a useful mathematical

and logical inference toolinference tool on decision making and

knowledge management.


Decisional Trust

2.2. Reflexive OntologiesReflexive Ontologies If a knowledge structure such as the Decisional DNA is enhanced

with the capabilities of ontology based technology, its performance

is increased in terms of two characteristics:

complementary inference capabilitiesinference capabilities; and

share abilitiesshare abilities given by the semantic annotation meta-languages in

which ontologies are transmitted.

Adding heavier semantics, logic, and expressiveness to the

Decisional DNA resulted in an OWL decisional Ontology, which is

broaden even more with the capabilities of a Reflexive Ontology

profiting in performance for its additional propertiesadditional properties.


Decisional Trust

3.3.SecuritySecurityFinally, all this knowledge is boosted

with security and signature security and signature

technologiestechnologies transforming such

knowledge into Decisional TRUSTDecisional TRUST

within the Semantic Web.


Decisional TrustO

nto

log

yO

nto

log

y

VariablesVariables

Rules Rules Constraints Constraints FunctionsFunctions

SOEKSSOEKS

Decisional ChromosomeDecisional Chromosome

Decisional DNADecisional DNA RORO

TrustTrust

Se

curity

Se

curity

KnowledgeKnowledge


THANK YOUTHANK YOU

content

Documents