AI in Knowledge Management Professor Robin Burke CSC 594.

AI in Knowledge Management

Professor Robin Burke

CSC 594

Outline

Introduction to the class Overview

Knowledge managementAICase-based reasoning

Objectives

Content Explore AI applications in knowledge

management• specifically case-based reasoning

SkillsReading research literatureBuilding an informal knowledge base

Course design

Seminar formatstudent presentationsin-class exercises

Attendance VERY IMPORTANT! Reading VERY IMPORTANT!

Reading

Two main readings each weekcase studyresearch article

Admission ticket1-2 page reaction paperwhat did you find interesting?a discussion question

Assessment

Presentations – 40% two presentations / student 1 case study 1 research paper

Participation – 50% course librarian discussion

Final Project – 10% more later

Typical class session

Case study30 min. presentation15 min. discussion

Research paper30 min. presentation15 min. questions

Librarian’s reports Group exercise

Artificial intelligence

The subfield of computer science concerned with the concepts and methods of symbolic inference by computer and symbolic knowledge representation for use in making inferences.

AI can be seen as an attempt to model aspects of human thought on computers. It is also sometimes defined as trying to solve by computer any problem that a human can solve faster.

-- FOLDOC

Knowledge management

Knowledge management involves the acquisition, storage, retrieval, application, generation and review of the knowledge assets of an organization in a controlled way.

-- I. Watson

Example: oil industry

old model own oil wells pump oil sell it

problem how to grow when there’s no more wells to

own? volatility of oil market low margins for commodity products high costs

Example: cont’d

solution: reconceptualize businessoilfield expertise

benefitseveryone needs know-howexpertise is always valuable

Hierarchy of knowledge

Knowledge expert analysis synthesis integration with experience

Information reports on data summarization

Data recorded information

The world stuff happens

Knowledge assets

Usually intangiblein worker’s heads

How to make experience explicit?not just what?but also why, how, and why not?

AI + Knowledge Management

Model aspects of human thought on computers

Which aspects?the storage and use of experience

What sub-field of AI studies this?case-based reasoning

Problem-solving

One of the first two areas tackled by AI researchother is natural language

How do we solve problems?researchers looked at logic puzzles

and problems of robot control

Rule-based reasoning

What are the steps to the solution? problem situation desired result

Forward-chaining reason forward from the problem

Backward-chaining reason backward from the desired state

Build up large rule bases also control knowledge

Case-based reasoning

An alternative to rule-based problem-solving

“A case-based reasoner solves new problems by adapting solutions used to solve old problems”

-- Riesbeck & Schank 1987

Paradox of the expert

Experts should have more rulescan solve more problemscan be much more precise

But experts are faster than noviceswho presumably have fewer rules

What does experience provide if it isn’t just “more rules”?

Problems we solve this way Medicine

doctor remembers previous patients especially for rare combinations of symptoms

Law English/US law depends on precedence case histories are consulted

Management decisions are based on past experience

Financial performance is predicted by past results

Retain Review

Adapt

Retrieve

Database

NewProblem

Similar

SolutionSolution

CBR Solving Problems

CBR System Components

Case-base database of previous cases (experience) episodic memory

Retrieval of relevant cases index for cases in library matching most similar case(s) retrieving the solution(s) from these case(s)

Adaptation of solution alter the retrieved solution(s) to reflect differences

between new case and retrieved case(s)

R4 Cycle

REUSEREUSEpropose solutions from retrieved cases

REVISEREVISEadapt and repair

proposed solution

CBRCBR

RETAINRETAINintegrate in

case-base

RETRIEVERETRIEVEfind similar problems

CBR Assumption New problem can be solved by

retrieving similar problemsadapting retrieved solutions

Similar problems have similar solutions

?

SSS

SS S

SS S

PP

PPPP

P

PP

X

AI in Knowledge Management

Apply the CBR model to the organization rather than the individualRetain the experience of the firmApply it in new situationsDo this in a consistent, automated

way

How to do this?

Very situation-specific What is a case? What counts as similar? What do you need to know to adapt

old solutions? How do you find and remove obsolete

cases?

CBR Knowledge Containers

Cases Case representation language Retrieval knowledge Adaptation knowledge

Cases

Contentslesson to be learnedcontext in which lesson applies

Issuescase boundaries

• time, space

Case representation language

Contentsfeatures and values of

problem/solution Issues

more detail / structure = flexible reuseless detail / structure = ease of

encoding new cases

Retrieval knowledge

Contentsfeatures used to index casesrelative importance of featureswhat counts as “similar”

Issues“surface” vs “deep” similarity

Nearest Neighbour Retrieval

Retrieve most similar k-nearest neighbour

k-NN Example1-NN5-NN

How do we measure similarity?

Can be strictly numericweighted sum of similarities of

features“local similarities”

May involve inferencereasoning about the similarity of items

Adaptation knowledge

Contentscircumstances in which adaptation is

neededhow to modify

Issuesrole of causal knowledge

• “why the case works”

Learning Case-base

inserting new cases into case-base updating contents of case-base to avoid mistakes

Retrieval Knowledge indexing knowledge

• features used• new indexing knowledge

similarity knowledge• weighting• new similarity knowledge

Adaptation knowledge

What this class is about

We will study examples of KM-related CBR applications

We will study CBR technology and research

Next week

Case study R. Burke & A. Kass (1994) "Tailoring

Retrieval to Support Case-Based Teaching." Proceedings of the 12th Annual Conference on Artificial Intelligence.

Research A. Aamodt & E. Plaza (1994) "Case-based

reasoning: Foundational issues, methodological variations, and system approaches." AI Communications, 7:39-59

Administrativa

Sign up for presentations Sign up for librarian slots