Knowledge Engineering

Lecture delivered by Dr S.Natarajan Professor , Dept of ISE, PESIT,Bangalore Session Chair for National Conference on Optimization of IT Oxford College of Engineering, Bangalore on February 17,2010

1. Background What Is Knowledge?

Data
A Record of a Change of State
1840KL0617

Information
Data organized with a purpose. A message
The flight from Delhi leaves at 18:40 hours

Knowledge
Literally what people know
thats not a good flight; Often busy and delayed

Knowledge engineering
process of

eliciting, structuring, formalizing, operationalizing

information and knowledge involved in a knowledgeintensive problem domain, in order to construct a program that can perform a difficult task adequately

Introduction

Ontology

An ontology is a formal representation of a set of concepts within a domain and the relationships between those concepts. It is used to reason about the properties of that domain, and may be used to define the domain In theory, an ontology is a "formal, explicit specification of a shared conceptualisation. An ontology provides a shared vocabulary, which can be used to model a domain that is, the type of objects and/or concepts that exist, and their properties and relations Ontologies are used in artificial intelligence, the Semantic Web, systems engineering, software engineering, biomedical informatics, library science, enterprise bookmarking, and information architecture as a form of knowledge representation about the world or some part of it. The creation of domain ontologies is also fundamental to the definition and use of an enterprise architecture framework.

Introduction

How are agents built

Intelligent Agent Domain Expert
Dialog Programming

Knowledge Engineer

Inference Engine

Knowledge Base

Results

A knowledge engineer attempts to understand how a subject matter expert reasons and solves problems and then encodes the acquired expertise into the agent's knowledge base. The expert analyzes the solutions generated by the agent (and often the knowledge base itself) to identify errors, and the knowledge engineer corrects the knowledge base.
G.Tecuci, Learning Agents Laboratory

Problems in knowledge engineering

complex information and knowledge is difficult to observe experts and other sources differ multiple representations:
textbooks graphical representations heuristics Skills A study carried out in 1989 showed that the main reason why expert systems were not being used was an insufficiency of methods for development, especially in the construction of knowledge bases, e.g. the transfer of expertise.

Introduction

Importance of proper knowledge engineering

Knowledge is valuable and often outlives a particular implementation

knowledge management

Errors in a knowledge-base can cause serious problems Heavy demands on extendibility and maintenance

changes over time

Introduction

A Short History of Knowledge Systems

general-purpose search engines (GPS)

first-generation rule-based systems (MYCIN, XCON)

emergence of structured methods (early KADS)

mature methodologies (CommonKADS)

1965

1975

1985

1995

=> from art to discipline =>

Introduction

First generation Expert Systems

shallow knowledge base single reasoning principle uniform representation limited explanation capabilities

reasoning control

operates on

knowledge base

Introduction

Main phases of the agent development process

Defining problem to solve and system to be built: requirements specification Understanding the expertise domain Feedback loops among all phases

Choosing or building an agent building tool: Inference engine and representation formalism

Development of the object ontology

Development of problem solving rules or methods

Refinement of the knowledge base

G.Tecuci, Learning Agents Laboratory

Investigated solution
Develop a computer system that incorporates the expertise of people familiar with spill detection and containment (i.e. a knowledge-based system, expert system or agent).

A knowledge engineer is assigned the job of building the system. The knowledge engineer becomes familiar with the problem and the domain. The knowledge engineer finds an expert on the subject who agrees to collaborate in building the system.
G.Tecuci, Learning Agents Laboratory

Elicitation of experts conception of a domain

By eliciting the expert's conception of his/her expertise domain we mean determining which concepts apply in the domain, what they mean, what is their relative place in the domain, what are the differentiating criteria distinguishing the similar concepts, and what is the organizational structure giving these concepts a coherence for the expert.

G.Tecuci, Learning Agents Laboratory

Elicitation methodology
(based primarily on Gammack, 1987)
1. Concept elicitation: methods (elicit the concepts of the domain i.e. an agreed vocabulary) 2. Structure elicitation: the card-sort method (elicit some structure for the concepts)

3. Structure representation (formally represent that structure in a semantic network) 4. Transformation of the representation (transform the representation to be used for some desired purpose)
G.Tecuci, Learning Agents Laboratory

Transfer View of KE

Extracting knowledge from a human expert

mining the jewels in the experts head expert is asked what rules are applicable translation of natural language into rule format

Transferring this knowledge into KS.

Introduction

14

Problems with transfer view

The knowledge providers, the knowledge engineer and the knowledge-system developer should share

a common view on the problem solving process and a common vocabulary

in order to make knowledge transfer a viable way of knowledge engineering

Introduction

15

Rapid Prototyping

Positive

focuses elicitation and interpretation motivates the expert (convinces management) large gap between verbal data and implementation architecture constrains the analysis hence: distorted model difficult to throw away

Negative

Introduction

16

Methodological pyramid

case studies application projects CASE tools implementation environments life-cycle model, process model, guidelines, elicitation techniques graphical/textual notations work sheets, document structure model-based k nowledge engineering reuse of k nowledge patterns

use t ools met hods t heory world view

feedback

Introduction

17

World view: Model-Based KE

The knowledge-engineering space of choices and tools can to some extent be controlled by the introduction of a number of models Each model emphasizes certain aspects of the system to be built and abstracts from others. Models provide a decomposition of knowledgeengineering tasks: while building one model, the knowledge engineer can temporarily neglect certain other aspects.

Introduction

18

CommonKADS principles

Knowledge engineering is not some kind of `mining from the expert's head', but consists of constructing different aspect models of human knowledge The knowledge-level principle: in knowledge modeling, first concentrate on the conceptual structure of knowledge, and leave the programming details for later Knowledge has a stable internal structure that is analyzable by distinguishing specific knowledge types and roles.

Introduction

19

KADS
Knowledge Acquisition and Documentation Structuring (KADS) is a structured way of developing Knowledge-Based Systems (KBS)(Expert Systems). It was developed at the University of Amsterdam as an alternative to an evolutionary approach and is now accepted as the European standard for KBS Its components are: A methodology for managing knowledge engineering projects. A knowledge engineering workbench. A methodology for performing knowledge elicitation. KADS was further developed into CommonKADS

Introduction

20

KADS (contd)

Knowledge Based Systems Analysis and Design Support (KADS) originating in the European ESPRIT project P1098 75 men-years of work, was one of the most highly developed KBs (Knowledge Based Systems) in the early 90s.

Introduction

21

KADS (contd)

This pioneering method provides two types of support for the production of KBs in an industrial approach: firstly, a lifecycle enabling a response to be made to technical and economic constraints (control of the production process, quality assurance of the system,...), and secondly a set of models which structure the production of the system, especially the tasks of analysis and the transformation of expert knowledge into a form exploitable by the machine.
22

Introduction

CommonKADS theory

KBS construction entails the construction of a number of models that together constitute part of the product delivered by the project. Supplies the KBS developer with a set of model templates. This template structure can be configured, refined and filled during project work. The number and level of elaboration of models depends on the specific project context.

Introduction

23

CommonKADS Model Set

Context Organization Model Task Model Agent Model

Concept

Knowledge Model

Communication Model

Artefact

Design Model
24

Introduction

Model Set Overview (1)

Organization model

supports analysis of an organization, Goal: discover problems, opportunities and possible impacts of KBS development. describes tasks that are performed or will be performed in the organizational environment describes capabilities, norms, preferences and permissions of agents (agent = executor of task).

Task model

Agent model

Introduction

25

Model Set Overview (2)

Knowledge model

gives an implementation-independent description of knowledge involved in a task. models the communicative transactions between agents. describes the structure of the system that needs to be constructed.

Communication model

Design model

Introduction

26

Principles of the Model Set

Divide and conquer. Configuration of an adequate model set for a specific application. Models evolve through well defined states. The model set supports project management. Model development is driven by project objectives and risk. Models can be developed in parallel.

Introduction

27

Models exist in various forms

Model template

predefined, fixed structure, can be configured objects manipulated during a project. versions of a model instance can exist. separate instances can be developed example: ''current'' and ''future'' organization

Model instance

Model versions

Multiple model instances

Introduction

28

The Product

Instantiated models

represent the important aspects of the environment and the delivered knowledge based system. information not represented in the filled model templates (e.g. project management information)

Additional documentation

Software

Introduction

29

Roles in knowledge-system development

knowledge provider knowledge engineer/analyst knowledge system developer knowledge user project manager knowledge manager

N.B. many-to-many relations between roles and people

Introduction

30

Knowledge provider/specialist

traditional expert person with extensive experience in an application domain can provide also plan for domain familiarization

where would you advise a beginner to start?

inter-provider differences are common need to assure cooperatio

Introduction

31

Knowledge engineer

specific kind of system analyst should avoid becoming an "expert" plays a liaison function between application domain and system

Introduction

32

Knowledge-system developer

person that implements a knowledge system on a particular target platform needs to have general design/implementation expertise needs to understand knowledge analysis

but only on the use-level

role is often played by knowledge engineer

Introduction

33

Knowledge user

Primary users

interact with the prospective system

are affected indirectly by the system

Secondary users

Level of skill/knowledge is important factor May need extensive interacting facilities

explanation consider attitude / active tole

His/her work is often affected by the system

Introduction

34

Project manager

responsible for planning, scheduling and monitoring development work liaises with client typically medium-size projects (4-6 people) profits from structured approach

Introduction

35

Knowledge manager

background role monitors organizational purpose of

system(s) developed in a project knowledge assets developed/refined

initiates (follow-up) projects should play key role in reuse may help in setting up the right project team

Introduction

36

Roles in knowledge-system development

knowledge manager defines knowledge strategy initiates knowledge development projects facilitates knowledge distribution

knowledge provider/ specialist

elicits knowledge from elicits requirements from

knowledge engineer/ analyst manages project manager

validates

delivers analysis models to KS uses knowledge user designs & implements knowledge system developer manages

Introduction

37

Terminology

Domain

some area of interest

banking, food industry, photocopiers, car manufacturing

Task

something that needs to be done by an agent

monitor a process; create a plan; analyze deviant behavior

Agent

the executor of a task in a domain

typically either a human or some software system

Introduction

38

Terminology

Application

The context provided by the combination of a task and a domain in which this task is carried out by agents The particular area of interest involved in an application The (top-level) task that needs to be performed in a certain application

Application domain

Application task

Introduction

39

Terminology

knowledge system (KS)

system that solves a real-life problem using knowledge about the application domain and the application task knowledge system that solves a problem which requires a considerable amount of expertise, when solved by humans.

expert system

Introduction

40

Knowledge engineering

GDS
600,000 Travel Agents use the GDS to find and book hotels & flights Used by IATA approved Travel Agents world wide
40,000 Internet Distribution Systems (IDS) may use GDS * Expedia, TraveloCity etc * IDS May pull content from GDS giving a single point of control for multiple channels This very powerful feature is however being depreciated as IDS opt for direct contracts with hotels Global Exposure!!

Knowledge engineering

GDS
Started By American Airlines to let travel agents book flights (1964*) Worlds airlines joined. All system networked together Expanded to include hotels (1988 chains via Thisco switch)
Travel agents search for hotels using a secure computer terminal connected to one of the 4 GDS channels - Amadeus Galileo Sabre Worldspan Each GDS displays your hotels current rates and availability 300 million reservation per month

The reservation is sent to the hotel. It is immediately available to the travel agent via the GDS and to the hotel via the GDS agent CRS (Generares and partner system (arcRes) 1964* SABRE: Semi-Automated Business Resrch Environment. The larges non government database in the world

Knowledge engineering

GDS information is also available to display and book on 1,000s of travel sites like Travelocity

Knowledge engineering

GDS NEWORK
600,000 Travel Agents use the GDS to find and book hotels & flights 40,000 Internet Distribution Systems (IDS) may use GDS *

GDS-CRS Rep Eg: Reserv, Unirez, Utell (Pegasus) Synix.. etc and Genares Interfaced to your back office management/accounting and front office reservation and web

Knowledge engineering

GDS BENEFITS

A Global Bookings Channel

Global Marketing

exposing you to the global travel market

Administration

manages rates on multiple channels - integrates with backoffice, front office

Knowledge engineering

GENARES GDS-CRS
Combining over 35 years of experience in the hospitality reservation technology industry, GENARES has developed the first truly integrated third party central reservation system for the twenty-first century

Complete open specification and interfaces for 3rd party integration, using XML format
The youngest and fasting growing GDS integration company Highly recommended by existing clients

Knowledge engineering

GENARES RECOMENDATIONS
Hi Ian, GenaRes is one of the best GDS providers Ive worked with within my 27 year in the industry! In over two years now, our volume of business percentage has grown through bookings via the GDS. I have been assigned my own support account manager and we have over the years established such a great working relationship. She is always here for me and also when Im out of the office she is the contact person regarding any issues with loading rates, opening/closing the system etc. Although the property can control basically everything own their own through the easy GenaRes system; my account manager has always proven to be our second hand person. This is what I call a great support team!!

Thanks Ian and Ill be on my way to Barbados to pick up that Rum Punch!! Call on me anytime and be well. Best Regards, Clayton C.Channing Reservations/Revenue Manager- flatotel.com 646-756-7952

And many more

Knowledge engineering

Genares

- full featured CRS

Reporting

Agents Commission Tracking Integration

RoomMaster, IQ Ware, RSI, AutoClerk, * Opera,* Check Inn, PMS Solutions/Innkeeper, Expedia Quick Connect, ezyield

& arcRes

Knowledge engineering

Integrated with AXSES ArcRes for fast, easy setup & management

RoomMaster, IQ Ware, RSI, Check Inn Expedia Quick Connect AutoClerk*, Opera,* PMS Solutions/Innkeeper*
* In development

Knowledge engineering

GDS and arcRes easy to use & powerful .

Easy navigation on your arcRes home page

. Just a click away

Knowledge engineering

arcRes GDS Setup and Load easy as 123

.

Knowledge engineering

arcRes - Register GDS with a Click

Registering switch and contract online for easy access and process
.

Automatically generate letter and contracts with a click no paperwork!

Knowledge engineering

Load GDS rates and Content using arcRes info .

Takes all info in arcRes and formats it for GDS

Allows you to save an edit for as long as you want

. Saves time and improves content

We dont know of a GDS that is as easy and as cost effective to setup

Knowledge engineering

Generes GDS Marketing Key Partners

40,000 travel Sites (IDS) May pick up GDS content

Participation in worldwide trade shows NBTA HITEC HEDNA All (GDS) conferences WTM (World Travel Mart), attendee only ITB, attendee only ResExpo RFP consortia participation Private GDS chain level branding programs

Knowledge engineering

GDS as Channel Manager

40,000 travel Sites (IDS) May pick up GDS content

IDS pull from GDS (sometimes exclusively *) Some have Direct (net rates) contacts Many do both IDS Net Rates Contract 25 - 35% Barbados Hotels pay for placement (loss of contract = loss of position)

Being on GDS may not eliminate need to manage direct contract (net rates) with companies like Expedia

Genares offer IDS Direct like Expedia Quick connect

* priceline

Marketing Constraints

Knowledge engineering

AFORDABLE GDS .. 5% GDS rep fee

GDS Setup $150 - arcRes time saver load $ 50 -Pass through -GDS Rep - minimum monthly maintenance fee is - ODD monthly connection fee is

$ 5.75 5% $25.00 $25.00

Website arcRes Bookings (i) Dynamic rates arcRes . arcRes channel Management option
Agent Commission (TACS) per reservation to GDS per reservation to Perot Systems

$0 $ 75 $ 2.00

$ 0.45 $ 0.55

Knowledge engineering

AFORDABLE GDS .. Just Got Better!!

Combined with arcRes for

- one of a kind marketing - full service - easy GDS setup - savings
.

Knowledge engineering

arcRes

Interactive services Increase bookings by 25%

.

Reservations Booking engine

BookingsBarbados channel Search engine Comparison Shopping Dynamic packages * Bookable Advertising! GDS Global Distribution Themes Specials channel Pre-made packages Groups management Dynamic rates & accom. CMS Affiliate marketing

750 3000 / 1500 3750

Knowledge engineering

Dynamic Bookable Advertising

rates quotes .

Full shopping cart cost and compare **** huge increase in bookability***

Knowledge engineering

Dynamic Bookable Advertising

Cost Compare of all point select. Quotes, invoice, save, book

quotes .

Knowledge engineering

Dynamic Bookable Advertising

NOT Bookable xx

Linked to website if advertising

- no rates - no quotes

- no shopping
- no consistent information

Knowledge engineering

Dynamic Bookable Advertising

everywhere!!

>>>>>>

Soon. A bookable map for every arcres advertisier.

Knowledge engineering

Advertise - Search Shop Buy . Blur >> The Perfect Storm element in the search-shop-buy triumvirate is undergoing a period of intense innovation, making each increasingly significant, yet interdependent. In fact, searching, shopping and buying once distinct terms describing different behaviors are blurring at a furious pace. Philip C. Wolf, President and CEO, PhoCusWright Inc. AXSES is there. We have already integrated advertising with all phases of the shopping cycle. This gives you complete flexibility in revenue and marketing models; including any mix of transaction, commission and subscription. Focus direct sales facilitating distribution Interactive advertising works! Travelers stay longer and use all options
Each

Knowledge engineering

arcRes-GDS PACKAGE

- twice as affordable

Pay gds setup/ we credit it and give 50 % off on balance

GDS setup Credit $300 PACKAGE Discount $700 Bookable-ads on Barbados.org RealHolidays Dynamic Packaging (RH)

BookingsBarbados (BB) ResortSearch, Comparison Shopping Quotes and Bookings Own Bookings engine (no commissions)

You cant be without it

Interactive advertising works! Travelers stay longer and use all options

Knowledge engineering

arcRes GDS PACKAGES -

combinations

Search, comparison shopping quotes, reservations, bookings Dynamic packaging, Bookable-ads

Search, comparison shopping, quotes, reservations, bookings Bookable-ads

Saves $1000
Applies full GDS costs $300 Full RH Cost $500 Plus 50% on remainder

Saves $450
Applies full GDS costs $300 Plus 50% on remainder

Non hosted - clients pay additional $250 setup and $250 pa 50%

$250

Knowledge engineering

Turn on the light

Bookable-Ads New Destination Channels

GDS exposure/bookings
.

Fully automated setup