Docsity
Docsity
Prepare for your exams
Prepare for your exams

Study with the several resources on Docsity

Earn points to download
Earn points to download

Earn points by helping other students or get them with a premium plan

Guidelines and tips
Guidelines and tips
Sell on Docsity
Sell on Docsity
Docsity AI
Docsity AI
Log inSign up

Prepare for your exams

Study with the several resources on Docsity

Find documents

Prepare for your exams with the study notes shared by other students like you on Docsity

Search for your university

Find the specific documents for your university's exams

Docsity AINEW

Summarize your documents, ask them questions, convert them into quizzes and concept maps

Explore questions

Clear up your doubts by reading the answers to questions asked by your fellow students

Earn points to download

Earn points by helping other students or get them with a premium plan

Share documents
download point icon20 Points

For each uploaded document

Answer questions
download point icon5 Points

For each given answer (max 1 per day)

All the ways to get free points
Get points immediately

Choose a premium plan with all the points you need

Study Opportunities

Choose your next study program

Get in touch with the best universities in the world. Search through thousands of universities and official partners

Community

Ask the community

Ask the community for help and clear up your study doubts

Free resources

Our save-the-student-ebooks!

Download our free guides on studying techniques, anxiety management strategies, and thesis advice from Docsity tutors

Knowledge Acquisition - Artificial Inteligence - Lecture Slides, Slides of Artificial Intelligence

In the class of Artificial Inteligence we learn the basic concept of programming, here are some major points discuss in these lecture slides which I shared with you:Knowledge Acquisition, Uncertainty, Knowledge Engineering Process, Unaware, Solving Strategies, Inconsistent Information, Expert Statements, Three Approaches, Conceptual Model, Issues

Typology: Slides

2012/2013

Uploaded on 04/23/2013

saritae
saritae 🇮🇳

4.5

(10)

101 documents

1 / 8

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
1
Knowledge Acquisition
and
Uncertainty in ES
Knowledge engineering process
(general)
Knowledge eng ineering process (based on Negnevitsky p 300)
Integration and m aintenance
Evaluation of system
Complete system de velopment
Prototype developm ent and testing
Data and knowledge acquisition
Problem assessm ent
Knowledge acquisition (KA)
challenges
• Experts often unaware of (or unable to
explain) their own knowledge and/or
problem solving strategies
• Experts can provide irrelevant, incomplete
or inconsistent information.
• Turning expert statements into a consistent
rule set can be challenging.
Three approaches
• Manual: Skilled and experienced knowledge
engineer (KE) interviews expert, makes rules,
builds prototype, etc. Very labor intensive.
• Automatic: ES queries expert directly, and
attempts to formulate rules. Requires expert who
can think a bit like a KE!
• Mixed: KE uses automatic tools to help smooth
the process, but maintains high level control.
Usually the most effective for big KBs.
Goal: conceptual model of expert’s
knowledge
• Nature of model depends on domain
• Must be comprehensible (or translatable) to
expert, so that s/he can agree/disagree
• Could be linguistic, graphical, tabular, or
other.
A bit of a ‘black art’, rather than a technology
or methodology!
Issues in KE (from [Hart, 1986, p 34)
(1)
• What are the inputs or problems?
• What are the outputs or solutions?
• What types of inputs cause difficulty for the
expert?
• How are the problems characterized?
• How are the solutions characterized?
• What sort of knowledge is used?
• How are problems or methods broken down into
smaller units?
Docsity.com
pf3
pf4
pf5
pf8

Partial preview of the text

Download Knowledge Acquisition - Artificial Inteligence - Lecture Slides and more Slides Artificial Intelligence in PDF only on Docsity!

Knowledge Acquisition

and

Uncertainty in ES

Knowledge engineering process

(general)

Knowledge engineering process (based on Negnevitsky p 300)

Integration and maintenance

Evaluation of system

Complete system development

Prototype development and testing

Data and knowledge acquisition

Problem assessment

Knowledge acquisition (KA)

challenges

  • Experts often unaware of (or unable to

explain) their own knowledge and/or

problem solving strategies

  • Experts can provide irrelevant, incomplete

or inconsistent information.

  • Turning expert statements into a consistent

rule set can be challenging.

Three approaches

  • Manual: Skilled and experienced knowledge

engineer (KE) interviews expert, makes rules,

builds prototype, etc. Very labor intensive.

  • Automatic: ES queries expert directly, and

attempts to formulate rules. Requires expert who

can think a bit like a KE!

  • Mixed: KE uses automatic tools to help smooth

the process, but maintains high level control.

Usually the most effective for big KBs.

Goal: conceptual model of expert’s

knowledge

  • Nature of model depends on domain
  • Must be comprehensible (or translatable) to

expert, so that s/he can agree/disagree

  • Could be linguistic, graphical, tabular, or

other.

A bit of a ‘black art’, rather than a technology

or methodology!

Issues in KE (from [Hart, 1986, p 34)

(1)

  • What are the inputs or problems?
  • What are the outputs or solutions?
  • What types of inputs cause difficulty for the

expert?

  • How are the problems characterized?
  • How are the solutions characterized?
  • What sort of knowledge is used?
  • How are problems or methods broken down into

smaller units?

Issues in KE (from [Hart, 1986, p 34)

(2)

  • What are the interrelationships between data items?
  • How important/accurate are data items?
  • Which data might be missing?
  • What assumptions does the expert make?
  • What constraints?
  • What sort of inferences?
  • How are concepts and hypotheses formed?
  • How do they relate?
  • How and when does the expert change beliefs?
  • What evidence suggests particular goals/concepts?
  • What are the causal relationships?
  • Are there any logical constraints?
  • Which problems are hard, easy, interesting etc?

Fact finding by interviews

  • KE records (preferably video) interview with

expert.

  • Goal is to get expert talking freely about the

problem

  • KE must analyze the explicitly stated facts and

rules, as well as the implicit (I.e. the way the

expert manipulates knowledge)

  • KE must ask specific questions, and elicit specific

answers

  • KE must get both general rules, and exceptions

KE questioning searches for: [Hart,

p58]

  • Interesting cases: gives exceptions and detailed

reasoning

  • Distinguishing goals: i.e. intermediate states of

belief, that are sought out (e.g. “first we need to

figure out X.”)

  • Forward or backward chaining by expert
  • Talk through: Expert is asked to ‘think aloud’

while solving real or hypothetical problems

  • Decision analysis.

Decision analysis

1. List all possible decisions

2. For each decision, list consequences

3. For each consequence, assess worth and probability

4. Calculate expected worth of each consequence

(worth * probability)

5. Expected worth of decision is sum of expected

worths of consequences

6. Select the decision that maximizes worth

Problem: Depends on quantitative estimates of worth

and probability! Still good interview tool.

Feedback

  • Linguistic (i.e. translation of rules)
  • Tabular
  • Conceptual/semantic graph
  • Interaction with system

Semantic network example













Conditional probability (3)

If A depends on many mutually exclusive

events:





   



   ^  





     

Conditional probability (4)

So (2) simplifies to:

  



&"^ ^ ^ ^ ^ ^ 

Conditional probability (5)

If A depends only on B and not B, then (3)

becomes:

     ^     ^ 

 "   

       

  

Bayesian reasoning

  • Instead of A and B, consider H (a hypothesis) and

E (evidence for that hypothesis).

  • Given E, what is the probability of H being true?
  • P(H|E) is referred to as the posterior probability of

H given evidence E.

  • Expert provides prior probabilities for p(H), p(not

H), as well as conditional probabilities p(E|H) and

p(E|not H).

Bayesian reasoning: multiple

hypotheses and evidences

For many hypotheses and evidences, we can

generalize (4):





Bayesian reasoning: multiple

hypotheses and evidences (2)

If we assume conditional independence among

different evidences, simplifies to:

  



       

In class exercise

Your expert gives you the following table of prior

and conditional probabilities:

P(E 3 |Hi) 0.4 0.3 0.

P(E 2 |Hi) 0.0 0.5 0.

P(E 1 |Hi) 0.6 0.2 0.

p(Hi) 0.25 0.35 0.

Probability i=1 i=2 i=

Hypothesis

Likelihood of sufficiency

Measure of an expert’s level of belief in H if

E is present.

Likelihood of necessity

Measure of an expert’s disbelief in H if E is not

present:

Prior odds

For simplicity of calculation, we use prior

odds (instead of prior probability):

 

   ^ 

   ^ 

Problems with the Bayesian

approach

  • Humans are not very good at estimating

probability!

  • In particular, we tend to make different assumptions when calculating prior and conditional probabilities
  • Reliable and complete statistical information often

not available.

  • Bayesian approach requires hypotheses to be

independent and exhaustive, and evidences to be

at least conditionally independent – often not the

case.

  • One solution: certainty factors

Certainty factors

Based on Ch 3 in Negnevitsky

Use of CFs (2)

For conciseness, can also assign CFs to

different consequents, e.g.:

  • If animal lays eggs THEN
    • animal is bird {CF 0.3}
    • animal is reptile {CF 0.2}

Note that CFs don’t need to add up to 1 –

leftover is “other”.

Use of CFs (3)

CFs are also applied to evidence, e.g.:

  • animal lays eggs {cf 0.6}
  • temperature is high {cf 0.9}
  • user likes red wine {cf 0} (i.e. user has

never tried red wine)

Represents reliability or availability of

evidence. Typically given by the user at run

time.

Propagation of CFs

For a single antecedent rule:

    ^ 

Single antecedent rule example

  • IF patient has toothache THEN problem is

cavity {cf 0.3}

  • Patient has toothache {cf 0.9}

What is the cf(cavity, toothache)?

   !!""   ^   

Propagation of CFs (multiple

antecedents)

For conjunctive rules:

    ^ 

Conjunctive example

  • IF patient has toothache AND patient has

prior cavities THEN problem is cavity {cf

  • Patient has toothache {cf -0.5}
  • Patient has prior cavities {cf 0.9}
  • What is the cf for “problem is cavity”?

          

Disjunctive example

  • IF patient has toothache OR patient has

prior cavities THEN problem is cavity {cf

  • Patient has toothache {cf -0.5}
  • Patient has prior cavities {cf 0.9}
  • What is the cf for “problem is cavity”?

  ^   ^   

Multiple rules affecting H

If the hypothesis is affected by several rules:

       (^)      (^) 

  (^)    (^) 

       ^      

! % 9 :

! 0 %;:

! %;:

Multiple rules example:

  • IF patient has toothache THEN problem is

cavity {cf 0.3}

  • IF patient has prior cavities THEN problem

is cavity {cf 0.7}

  • Patient has toothache (cf –0.5)
  • Patient has prior cavities (0.9)

Calculate cf for “problem is cavity”.

Bayesian vs certainty factors

  • Prob theory is ‘good math’, and works well

if statistical data is available and accurate

probability statements can be made

  • CF theory lacks formal mathematical

foundation, but better able the kind of

estimates an expert is likely to make of

probability

Exercise

  • Add CFs to the rules in the “lights out” rule

base.

  • Run through an example, and calculate

resulting CF