Košice, 10 February 2006 1

Experience Management based on Text Notes

The EMBET System

Michal Laclavik

Košice, 10 February 2006 2

Overview Applications Motivation, history Experience Management Approach Ontology Knowledge Model Modeling Methodology EMBET Architecture Algorithms Applications Conclusion

Košice, 10 February 2006 3

Collaboration among users Knowledge Sharing Recommendation

Representation of Experience or Knowledge Text Notes

Objectives, Application of the system

Košice, 10 February 2006 4

Motivation, problem area In Pellucid IST Project

Active Hint approach AH = action on resource(s) in context + explanation Many AHs in Pellucid:

See Text Note in this context because is useful In K-Wf Grid project

Need for sharing of expert knowledge Text Notes

Natural way for people If we are able to detect context of note - note can help

others better then other formalized knowledge People like to enter notes or memos to remind something to

themselves or others

Košice, 10 February 2006 5

State of the Art – Ontologies,

Knowledge

Ontologies Knowledge Representation OWL-DL compatible with Description

Logic Query and Storage Engines

available RDF, OWL, RDQL based

Application domain Knowledge Management (KM) is

the process through which organizations generate value from their intellectual and knowledge-based assets (Source: CIO Magazine)

Experience Management is special kind of KM – based on “lessons learned”

Characters

Data

Information

Knowledge

Actions

Syntax

Semantics

Pragmatics

Reasoning

(Bergman, 2002, Experience Management)

Košice, 10 February 2006 6

Research Challenges Experience (Knowledge) Management Text Processing Knowledge, Semantic, Ontologies Semantic Annotation Domain Models (Flood Prediction, Traffic

Simulation, …) User Interaction Knowledge Relevance, Problem detection

Košice, 10 February 2006 7

Experience Management Approach

Problem p Problem Space (P) In EMS Case-Lesson pairs (c, l)

Case Space (C) lesson space (L).

maps problem space to case space c = f(p)

Košice, 10 February 2006 8

General vs. EMBET Approach General EM Approach

Characterize a problem Transform the problem from the space P to the space C. Choose from the cases the most "useful" lesson from the case-lesson

pairs stored in the database Apply that lesson.

EMBET EM Approach User context detection from environment which describes problem P Our Model is described by ontology and Notes are stored with

associated context, which describes space C Notes represent learned lesson L which is associated with space C

(note context). The note context is matched with a user problem described by the detected user context. The user context is wider than the note context and as a result all applicable notes are matched and returned.

Applying the lesson is left to the user be reading appropriate notes.

Košice, 10 February 2006 9

Thesis Objectives Design of Generic Ontology model for

Experience management with extension for different application domains

Design of Modeling Methodology for extension of the Ontology based Knowledge Model for different applications

Design & Development of generic and customizable EMBET Experience Management software with the Ontology Knowledge Model

Evaluation of results on a real pilot operation

Košice, 10 February 2006 10

Used Methods and Methodologies

Knowledge management, system design Unified Modeling Language – UML CommonKADS Protégé as Tool for CommonKADS

Formal methods for describing ontology based models Description Logic Graph Ontology representation

Košice, 10 February 2006 11

Used Tools and Software

Protégé Ontology Editor Support for OWL ontology format Can be used as modeling tool

Jena – Semantic Web Framework for Java Support for OWL – best available OWL

API Support for RDQL model querying

Košice, 10 February 2006 12

Ontology Knowledge Model

Objective:Design of Generic Ontology model for Experience management with extension for different application domains

Košice, 10 February 2006 13

Ontology Knowledge Model

Based on Events, Resources, Actions, Actors, Context

Formally Described using Sets, Description Logic (compatible with OWL-DL), Graph Representation

Actor Context updating function/algorithm (Actor/User Environment State)

CAnew = fC(ea,CA

old) Resources updating function/algorithm

(result of fulfilled actor/user goals) RA

new = fR(CAnew,RA

old)

Košice, 10 February 2006 14

Modeling Methodology

Objective:Design of Modeling Methodology for extension of the Ontology based Knowledge Model for different applications

Košice, 10 February 2006 15

Modeling Methodology

Extending Model with Protégé Editor following CommonKADS models

Organizational or Environment Model Task Model Agent or Actor Model

Includes implementation of algorithms for context and resource updating

Results Ontology developed in Protégé which can be

exported in OWL format. Concrete Algorithms for each actor (often algorithms

are similar or same) which updates actors' context CA

new and resources RAnew.

Košice, 10 February 2006 16

Design & Development of EMBET System

Objective:Design & Development of generic and customizable EMBET Experience Management software with the Ontology Knowledge Model

Košice, 10 February 2006 17

Use Case

Display Note Enter Note Vote on

Relevance Enter Problem

Košice, 10 February 2006 18

Knowledge Cycle

1Matching User

Context or Problem

3 User take

Decisions or Actions

2 Displaying Relevant

Notes

4User feedback

and knowledge update

(4) Enter Note

(4) Pattern DetectionAnnotation Note with Context

(1, 4) Compare Current User ContextAnd Note Context

(1, 4) Displaying context List

(4) User approve and submit context of note

(2) Notes are displayed to user in detected user context

(4) Vote on note,Updating relevance

Košice, 10 February 2006 19

Architecture and GUI

XML messages

JSPPages

HTML Output

JSTL

XSLT Style Sheets CSS

EMBET GUI

RDFOWL

XML messages

KnowledgeNotes

FeedbackOn

knowledge

UpdateKnowledge

ContextUpdate

EMBET Core

KnowledgeNotes

Detection

ContextDetection

StoreKnowledge

Memory

Košice, 10 February 2006 20

Class Diagram

Memory Ontology Anotate Core EventHandler DetectContex

t

Košice, 10 February 2006 21

Algorithms Actor (User)

Context updating algorithm

CAnew = fC(ea,CA

old)

Resources (Notes) updating algorithm

RAnew = fR(CA

new,RAold)

Procedure updateContext(Event e) // e - new event in the system relevant to actorbegin actor = e.actor; //event is related to some actor (user) if e.action = Create or e.context is empty then begin //when a resource is added (like note) or context of

event is empty //we can add context of actor to the event while c = foreach(actor.context) do e.context += c; end else begin //if actor (user) performed some action or action

related to user happened //we need to update actor context based on received

event context remove(actor.context); while c = foreach(e.context) do actor.context += c; endend

Procedure updateResources(Actor actor)begin // unlink any notes of actor unlinkActorNotes(actor); while noteS = foreach(Note from Memory) do begin // checking defined note context elements matchAll = true; while c = foreach(noteS.context) and matchAll do begin if actor.context has not c matchAll = false; end //new note is added to actor if matchAll addProperty(actor.resource, noteS); endend

Košice, 10 February 2006 22

Annotation Text of note is

matched by regular expressions

Domain (Application) elements described in ontology model are detected

Košice, 10 February 2006 23

Applications

Objective:Evaluation of Results on real pilot operation.

Košice, 10 February 2006 24

Example of Use – Adding the Note

Košice, 10 February 2006 25

Example of Use – Definition of Problem

Košice, 10 February 2006 26

Conclusion and Future Work

Košice, 10 February 2006 27

Major Outcomes/Results EMBET is system for:

User Problem Definition Experience Management Collaboration Knowledge

Sharing

Košice, 10 February 2006 28

Conclusion and outlook Tested and Evaluated on:

Pellucid IST Project K-Wf Grid IST Project

Can be used also in non Grid application Intranet Systems CRM, ERP Systems which can communicate Context

Customization Domain Ontology Model (main resources in the

domain) Interface communicating context/problem of the user

Košice, 10 February 2006 29

Future work RAPORT APVT project (01/2005-12/2007): Research and

development of a knowledge based system to support workflow management in organizations with administrative processes

model and algorithms will be reused and extended Annotation of emails

K-Wf Grid EU 6FP RTD IST project (2004-2007) evaluation on more applications, improvement of context

detection

NAZOU SPVV Project (09/2004-11/2007): Tools for acquisition, organization and maintenance of knowledge in an environment of heterogeneous information resources

OnTeA semantic annotation

Košice, 10 February 2006 30

Thank you !

Michal Laclavik

Košice, 10 February 2006 31

Questions & Answers

Košice, 10 February 2006 32

Methodologies

Based on CommonKADS Protégé

Tied with Extensible model for discrete environments based on Events

Košice, 10 February 2006 33

Experience Management Systems

EM as in Bergman Book – methodology for EM application

EMBET Generic

Košice, 10 February 2006 34

Annotation

Annotation solutions – semantic tags

OnTeA - Mapping with existing individuals in ontology models