You are on page 1of 27

Types of Knowledge

Commonsense Knowledge
Commonsense knowledge is the general knowledge about
domain that an average person is expected to know.
Informed Commonsense Knowledge
Informed commonsense knowledge is knowledge of given
domain that is common to all experts within the domain
Domain Knowledge
This type of knowledge refers to a specific area of
application such as medical, engineering, education etc.
Meta Knowledge
Knowledge about knowledge. How the knowledge is stored
what is the type, how it can be used to solve problem etc.

kbs Chapter-2
Classifying Knowledge according to
its nature
Tacit Knowledge
Ability to interpret, perceive, communicate etc.
Implicit Knowledge
Knowledge about systems, strategies, vision,
policies etc.
Explicit Non-codified Knowledge
Knowledge about procedures, instructions, history
of processes etc.
Explicit Codified Knowledge
Documentation, books, reports, directories etc.
kbs Chapter-2
Desirable Characteristics of Knowledge
Naturalness: Ease of representing knowledge in its
native form.
Transparency: Ease of identifying stored
knowledge of knowledge.
Adequacy and Completeness: Ability to contain all
components required to solve the problem.
Modularity: Modular representation facilitates
reuse of knowledge and hence reduce cost

KBS Chapter-2
Desirable Characteristics of Knowledge ..
Usefulness: Extent to which knowledge is useful to
solve a problem in the domain.
Explicitness: Ease of representing the knowledge
directly.
Ease of Operation, easy to access and efficient:
Ease of Obtaining, applying the knowledge to
problem solving and analyzing the results.

kbs Chapter-2
Components of Knowledge ….
The detail components of the knowledge are
1. Facts: It represents sets of raw observation, alphabets
symbols or statements
Eg:
 The Earth moves around the Sun.
 Every car has a Battery.
2. Rules: Rules encompasses conditions and actions, which
are also known as antecedents and consequences.
Eg:
 If there is a daylight, then the Sun is in the sky.
 If it does not start, then check the battery and the fuel.
kbs Chapter-2
Components of Knowledge…
3.Heuristics:
It is the way to represents problem- solving experience
within the knowledge base. They are solutions that an
expert in a similar situation. They do not give a guarantee
of success as, it is not developed from a standard formula
or process..
Eg
• If there is total eclipse of the Sun, there is no daylight,
even though the Sun in the sky.
• If it is a rainy season and a car was driven through
water, your silencer would have water in it, so it may
not start.

KBS Chapter-2
Basic Structure of Knowledge
Based System
A knowledge based system is the software containing a
source of knowledge called a knowledge base.
It is a mechanism for referring to existing knowledge and
generating new knowledge, that is inference, reasoning and
explanation.
1.Knowledge Base
2.Inference Engine
3.Self Learning
4.Reasoning
5.Explanation

KBS Chapter-2
General Structure of Knowledge-Based Systems
Knowledge base is a repository of domain
knowledge and metaknowledge.

Inference engine is a software program that


infers the knowledge available in the
knowledge base.

Enriches the
system with
Knowledge Inference
self-learning
base engine
capabilities

Explanation
Working Memory
and Self-
reasoning learning

Provides
explanation and User interface
reasoning Friendly
facilities interface to
users working
in their native
language

Figure : General structure of KBS


General Structure of Knowledge-Based Systems
User Interface
It provides a friendly interface to users working in their native
language
Knowledge-base
Knowledge base is a repository of domain knowledge and
meta-knowledge
Inference Engine
Inference engine is a software program that infers the
knowledge available in the knowledge base
Self Learning
Enriches the system with self-learning capabilities that
allows it learn from the environment and cases it solves
Explanation and Reasoning
Provides explanation and reasoning facilities. Explains, how
the problem is solved

KBS Chapter-2
Inference Engine
An inference engine is a software program that refers the existing
knowledge edge , manipulates the knowledge according to need and
make decisions about action to be taken.
It utilizes pattern matching and search techniques for
conclusions.
It examines the existing facts and rules and update the facts when
possible.
Decides the order of inferences.
Common rules for deriving new facts from rules and known facts.
 Modus Ponens
 Modus Tollens

KBS Chapter-2
Inference Engine…
Modus Ponens is a common inference strategy.
Frame work:
The rule states that when A is known to be true and a rule
states “ if A, then B” it is valid to conclude that B is true.
Ex:
Rule: IF the teacher is present in the class
THEN students must present in the class.
Given fact: The teacher is present and teaching in the class
on the given day.
Conclusion: Students must be present in the class.

12 KBS Chapter-2
Inference Engine…

Modus Tollens .
Frame work:
When B is known to be false and a rule states “ if A, then B”
then A is false.
Example:
Rule: IF the teacher is present in the class
THEN students must present in the class
Given fact: Students were absent on the given day.
Conclusion: The teacher must be absent, which is a new fact.

KBS Chapter-2
Forward Chaining(FC):
 The input of the working memory triggers rules for which
conditions match the new data and constraints.
 These rules then perform their actions, that may add new
data to memory, thus triggering more rules and so on (data
directed inference).
 The IE with FC searches the inference rules until it finds
one where an “IF” clause is known to be true.
 When found, it can conclude or infers, THEN clause,
which results in the additional information

KBS Chapter-2
Forward Chaining(FC): Algorithm

Step 1:Consider the initial facts and store them in working memory
of the knowledge base
Step 2:Check the antecedents part(left side) of the production
rules.
Step 3: If all the conditions are matched, fire the rule (execute the
right side)
Step 4: If there is only one rule, do the following:
Step 4.1: Perform necessary actions.
Step 4.2: Modify working memory and update facts.
Step 4.3: Check for new conditions
Step 5: If more than one rule is selected, use the conflict resolution
strategy to select the most appropriate rules and go to Step 4.
Step 6: Continue until an appropriate rule is found and executed

KBS Chapter-2
Backward Chaining(BC):

It uses piece of data or a hypothesis.


It searches for rules whose conclusions mention
this data.
System test the conditions, before using the rules.
This may entail discovering the value of more
pieces of data..(goal- directed inference).
 Inferences are not performed until the system is
made to prove a particular goal or question.

KBS Chapter-2
Backward Chaining(BC): Algorithm

Step 1:Start with a possible hypotheses, say H.


Step 2: Store the Hypothesis H in working memory, along with the
available facts. Also consider a rule indicator, R , set to Null.
Step 3: If H is in the initial facts, the hypothesis is proven. Go to
step 7.
Step 4: If H is not in the initial facts find a rule, say , R, that has a
descendent(action) part mentioning the Hypothesis.
Step 5: Store R in working memory.
Step 6: Check conditions of R and match with the existing facts.
Step 7: If matched, then fire the rule R and stop. Otherwise,
Continue to step 4.

KBS Chapter-2
Forward Versus Backward Chaining
Data directed inference , or forward chaining, is appropriate
when:
 There exits sufficient information about an environment to
conclude a final decision.
 There is a single initial state.
 It is difficult to form a goal to verify.
Goal- directed inference , or backward chaining, is appropriate
when:
 The goal is given or obvious.
 Environmental constraints or data are not clear.
 Relevant data must be acquired during the inference process.
 A large number of applicable rules exist.

KBS Chapter-2
Forward Versus Backward Chaining
Forward Chaining starts with some facts and applies rules to
find all possible conclusions and goal state is unpredictable
and unimportant.
It is mainly used in quality improvement phase in software
development.

Backward Chaining starts with the desired conclusions and


works backward to find supporting facts
• If there are few goals and small number of facts, BC is more
suitable as it avoids unnecessary efforts input large amount
of data.
• BC is good for diagnostics and classification tasks but not
preferable for planning , design, process monitoring
KBS Chapter-2
Forward Chaining Technique Example
Initially working memory has following assertions.
1. runny nose
2.temperature = 104 f
3.headache
4. cough
Rules in the Rule base are:
• R1: if (nasal congestion and viroses), then diagnose (influenza) exit
• R2: if (runny nose) then assert (nasal congestion )
• R3:if (body aches), then assert (itchiness)
• R4:if (temp>100), then assert (fever)
• R5:if (headache), then assert (itchiness)
• R6:if (fever and itchiness and cough), then assert (viroses)

KBS Chapter-2
Forward Chaining Technique
Example…..
The execution is as follows:
First R2 fires, so (nasal congestion) is added to
working memory.
R4 fires, so (fever) is added to working memory.
R5 fires, so (itchiness) is added to working memory.
R6 fires, so (virosis) is added to working memory.
R1 fires, diagnosing the ailment as (influenza) and
exit.

KBS Chapter-2
Forward Chaining Technique Example…..
S.NO Rule Working Memory Remark
Sequence
Before Execution
runny nose
Temperature = 104 f
headache
cough
After Execution
1 R2 nasal congestion
2 R4 fever
3 R5 itchiness
4 R6 virosis
5 R1 Action: Diagnose the
ailment as Influenza and
KBS Chapter-2
exit.
Backward Chaining Technique Example…
The execution is as follows:
First R1 fires, since goal is (influenza), goes with the conclusion of the
rule R1, new goals are created (nasal congestion) and (virosis).
 R2 fires, matching the goal (nasal congestion), new goal are created
(runny nose) , since the (runny nose) is in the memory it returns true.
 R6 fires, matching the goal (virosis), BC recursion takes place and new
goals are created (fever), (itchiness) and (cough)
 R4 fires, adding the goal (temperature> 100), as Temperature = 104 f is
in the memory it returns true.
 R3 fires , adding the goal (body aches), as (body aches), is not in the
memory it returns false.
 R5 fires adding the goal (headache), as headache is in the memory it
returns true.
 The goal cough is in the working, so that returns true.

KBS Chapter-2
Self Learning: It is scientific task that enables the
knowledge- based system to learn automatically
from the inference process, cases executed and
environment.
Reasoning: The capability and quality of a
knowledge based system depend upon the ability to
reason and explanations provided by experts. When
the knowledge –based system makes a decision,
especially when the system is not being told
explicitly how to perform the given task-
KBS Chapter-2
Applications Of Knowledge-Based Systems
1.Advisory systems
2.Health care and medical diagnostic systems
3.Tutoring systems
4.Control and monitoring
5. Prediction
6. Planning
7. Searching Larger Databases and Data Warehouses
8. Knowledge based Grid and semantic web
Advantages of KBS
Permanent documentation of knowledge
Cheaper solution and easy availability of
knowledge
Dual advantages of effectiveness and efficiency
Consistency and reliability
Justification for better understanding
Self-learning and ease of updates

KBS Chapter-2
Limitations of KBS
Partial Self-Learning.KBS can achieve human brain like
learning in a limited manner.
Creativity and Innovation: KBS system do not possess any
method to deal with creativity, innovation and common
sense.
Weak Support of Methods and Heuristics
Development Methodology
Knowledge Acquisition
Structured Knowledge Representation and Ontology
Mapping
Development of Testing and Certifying Strategies and
Standards for Knowledge-Based Systems

KBS Chapter-2

You might also like